0:00
[music] >> I feel multiple journalists on the horizon.
Search this video
0:08
>> [music] >> We're surrounded by journalists. Hold your position.
0:18
>> [music] >> Overnight success. The Plagiarizer 3000. Double click. Right. That's misinformation. Conflict of interest. Order a hit man. That's just wrong.
1:09
We're surrounded by journalists. Hold your position. Crop. Get out. Trust the experts. Buy Google. We are everywhere. Founder. Money.
1:38
>> [music] >> Buy Google.
1:43
I see multiple journalists on the horizon. Stand by. Experts agree.
1:52
>> [music] >> UAV online. Win. Double blades. Triple blades. Double kill.
2:15
>> [music] >> Fight back. Run. Come get some. Wait. Team death match. We are experts. Triple blades. Let's just run. Right.
3:07
Walking clearing order inbound. Come get some.
3:22
You're surrounded by journalists. Hold your position. Strike one. Strike two.
3:43
Activate golden retriever mode.
3:47
>> [music] >> Trust me, I got this.
3:57
>> All key clearing order inbound. Fire control. Extra fire.
4:18
I'll see more targets on the horizon. Head back. Found her.
4:33
You're watching TVP and today is Monday, March 9th.
4:36
We are live from the TVP and Ultra down.
4:39
Temple of technology, fortress of finance, ramp.
4:40
com, time is money, save money, use corporate cards, bill pay, accounting, and a whole lot more all in one place.
4:47
We have a fantastic show.
4:47
We were able to get Alex Epstein to come down to the Ultra down in person on short notice to get us up to speed on what's happening in the oil markets.
4:55
Of course, oil is all it's uh it's gushing all over the timeline, all over the Financial Times, all over the Wall Street Journal uh as part of the And Alex had a feeling that oil would be important >> Yes.
5:10
>> future when he wrote the book Fossil Future.
5:12
>> Yes, which we have right here.
5:14
Which if you haven't given it a read, go pick a copy up and stay tuned for him joining us at 1:10 today.
5:19
Let's pull up the linear lineup and tell you who we have coming on the show today.
5:24
Linear of course is the system for modern software development.
5:27
70% of enterprise workspaces on Linear using agents now. We have Dr.
5:32
Alex Wisner-Gross over on Eon Systems.
5:36
He went very viral over the weekend for connecting a biologically derived fruit fly connectome Yeah.
5:46
to MuJoCo physics simulated body.
5:50
Uh and uh we're going to be talking all about [laughter] simulating fruit flies Yes.
5:56
and what will come after that.
5:58
>> And we're also going over Microsoft Co-pilot Co-work, Julian and Sequoia. >> going to call CoCo. CoCo, okay.
6:03
Uh Disney might like a word.
6:06
They have uh IP in that space in particular, but we'll see.
6:09
>> they have it in the enterprise. I don't think so.
6:11
Um uh Julian's joining from Sequoia Capital.
6:14
Owen McCabe from from Intercom, our partner, is joining because they just raised a massive round.
6:19
clean 250 >> And Michelle Voles is launching Pax VC.
6:25
She's uh she left Andreessen Horowitz, I think about a year ago, and has been uh is now launching a new early-stage fund.
6:31
Uh anyway, let's go through the timeline and see how people are processing the news of the Strait of Hormuz and the oil price spikes.
6:38
Uh we >> Fox News this morning, they were saying just basically their advice to the to the captains was to man up and just send it. >> Yeah, send it. Send it. That seems crazy.
6:51
If you are on a shipping vessel, please please stay safe.
6:53
Um I'll read the actual quote.
6:56
Uh a guy, Brian, said, "If you want to diminish the Iranian threat, if you want to make sure this ends with complete capitulation, show some guts and go through that straight."
7:05
That seems very, very risky right now.
7:07
Stay safe out there if you are on a shipping vessel.
7:11
Um do whatever is responsible.
7:11
There are some crazy twists happening.
7:14
Uh so, apparently ships in the Gulf are declaring themselves as Chinese vessels to dodge attack.
7:22
This is from the Financial the Financial Times.
7:24
At least 10 vessels have changed transponder messages in an apparent attempt to avoid becoming targets.
7:30
Uh clutch of vessels trapped in the Gulf under enemy fire are adopting a tried and tested ruse to avoid attacks.
7:37
They're changing flags, uh using transponders to declare themselves to be Chinese.
7:42
Um there's there's always been a very odd uh like just like tug-of-war between how ships identify themselves because often for tax reasons they're bought in one country, operated by individuals from another company, but they fly a different flag to be able to go from one place to another, and it's all based on like the port systems.
8:04
I don't fully understand it, but it is very interesting.
8:06
Uh at least 10 ships over the past week have altered their destination signal to read Chinese owner, all Chinese crew, or Chinese crew on board.
8:13
About 1,000 ships are currently shut inside the Gulf and its immediate surroundings with a cumulative value of $25 billion.
8:19
And I'm going to I don't know if you've seen some of the the maps that show the strait where it looks like nothing is actually moving through. >> Yeah.
8:28
I think in actuality a number of the ships are actually turning off their transponders, so you can't see the movement. >> Yeah.
8:35
>> Uh because they're basically moving a little bit, going through the strait. Mhm.
8:38
And then uh turning back on again.
8:41
>> Well, there are there are plenty of uh outlets that are covering the uh the Iran conflict in detail.
8:45
Um we of course will be focused on the the business and technology impacts.
8:49
I wrote about uh what oil uh what what oil prices mean for the AI build-out and data centers.
8:57
It's just sort of interesting to dig into the deeper supply chain of artificial intelligence.
9:00
Um but there are some posts that we should go through around the oil story.
9:04
So, crude oil is five standard deviations above its 50-day moving average.
9:11
Statistically speaking, this occurs every 9,500 years.
9:15
So, the last time would have been about 6,000 years before Moses parted the Red Sea.
9:19
Imagine what that did to shipping in the area. Fanciful.
9:24
Um what else is going on?
9:27
Kobe Steel the Kobe Steel letter has been posting a lot about oil prices.
9:29
US oil prices are now expected to rise above $100 a barrel this month. That already happened.
9:35
Uh this was posted on March 6th uh with markets pricing a 66% chance that it happens.
9:40
The last time oil prices were above 100 was July of 2022.
9:45
>> Let's pull up this clip from Land Man. I haven't seen Land Man. Have you watched it? Is it good?
9:50
Let's uh Well, you want oil to live above 60 but below 90. Mhm.
9:56
And don't get me wrong, we're still printing money at 90, but gas gets up over 350 a gallon, it starts to pinch.
10:02
It hits a hundred, every product in America has to readjust its price.
10:08
$78 a barrel, that's about perfect.
10:08
You know, it brings enough profit to keep exploring, but it don't sting as much at the pump.
10:15
Unless of course you're in California.
10:15
I mean, they tax the out of it out there.
10:18
It could be $45 a barrel and it's still $4 at the pump.
10:21
I don't know how those son of do it out there. It's movie day.
10:27
>> 2020, a barrel of oil was worthless.
10:30
This place became a ghost town. And nobody's immune.
10:35
Kids have to quit college.
10:37
Trucks get sold or repossessed.
10:40
Uh we can pause it there.
10:40
We got we got the gist of it. >> Yeah.
10:42
Uh well, this is the largest How how old were you when you discovered that gas is really expensive in California not just because there's a lot of demand for it out >> I discovered it when I was filling up in Montana and it was like $2 a gallon and it's like $5 a gallon here. That was pretty wild.
10:57
Um the the first major gas price shock that I noticed was Hurricane Katrina. Mhm. my life.
11:03
I think [clears throat] it was in high school cuz I was too young >> I was talking about I was talking about just the fact realizing that it's so expensive because of the taxes that California puts on it.
11:15
>> Yeah, because it's nice to drive around here.
11:16
Put the top down, the weather's good, so they're like, "Yeah, you're going to pay you're going to pay more for the joys of driving an a an internal combustion engine car."
11:26
Now, uh the price in California is is aggressive.
11:28
I don't know the structure of the prices though.
11:30
Is it percentage-based or fixed?
11:32
Because that affects how much the price will move based on oil price shocks.
11:40
So, because if the price per barrel doubles, but the but the tax is flat, you don't feel it as much here as you do other places. So, I don't know.
11:49
Anyway, I'm pulling it up, but you can run through this.
11:52
>> supply shock by a factor of four.
11:52
So, the Hormuz blockade, which is current, uh 20 uh million barrels were lost in supply.
12:02
The Iranian Revolution in 1978 was 5. 6 million.
12:04
The Yom Kippur War embargoes in '73 was 4. 4.
12:07
The Iraq-Kuwait War was 4. 3 in 1990.
12:09
The Iran-Iraq War in 1980 was 4. 0.
12:12
And the Ukraine-Russia invasion in 2022, which is the last time that oil spiked over a $100 barrel, was one to three.
12:21
So, an absolutely huge supply shock.
12:24
And um I'm sure um it will have a lot of implications all over the economy.
12:31
Uh with triple-digit prices, here's what's going to happen now, says Policy Tensor.
12:35
Markets will scream when they open tomorrow.
12:37
VIX will surge to levels beyond what we saw in April.
12:42
The sell-off will continue for some time as intermediate as intermediaries shed risk. And the markets are red.
12:48
They have been screaming today.
12:53
Uh the VIX futures curve already has inverted, bid up by dealers looking for insurance.
12:56
This predicts a massive sell-off.
12:58
The pressure on this captured White House now uh the the pressure on this captured White House now beings in earnest.
13:08
That's kind of sort of oddly written.
13:09
Anyone can tell them that if this persists for very long, it will destroy the Trump presidency and gut the GOP for a generation.
13:15
The controlling factor here, as I have told you over and over again, is that the United States does not have the military means to reopen Hormuz.
13:22
Uh there's no military solution in sight.
13:25
This means that not only does Iran have the strategic upper hand now, it means that Iran enjoys the unambiguous strategic advantage.
13:31
All they need to do is keep the thing closed until he capitulates.
13:35
I put my neck out far to call this in advance, and someone told him yesterday I was in the minority, perhaps even a minority of one. No longer. I was correct.
13:46
He's just calling his shot.
13:48
The blob heads and scribes were incorrect in their assessment of the strategic situation, and now markets will price to reflect reality.
13:53
And we were uh we briefly touched on Scott Sumner's blog post on Substack at the end of Thursday show, maybe Friday show.
14:02
We didn't get a chance to read it.
14:04
I actually read through it uh earlier today. It's pretty interesting.
14:07
It basically makes the case you know, it has a very uh controversial title. So, freak out.
14:13
And and and the and the thought is that like he's doom posting.
14:17
Maybe it's about AI, maybe it's about uh this particular conflict.
14:18
He is more just reflecting on this dynamic between when the market freaks out, it acts as a moderating effect to policy.
14:26
And so, uh he gives a number of examples around like tariffs.
14:32
>> specifically the admin. Yeah.
14:32
Yeah, like like like the uh tariffs caused this massive circuit breaker 5% sell-off in the market, and then that was internalized and and very quickly adjusted, and there were a whole bunch of different carve-outs to sort of like create soft landings.
14:47
And so, he's actually and what what he's getting at is that after the fact, a lot of people say, "Look, you didn't need to freak out because the taco happened.
14:56
Trump always chickens out.
14:59
The the the actual proposed policy effect didn't go into didn't go into effect."
15:03
And his point is that uh "Well, it's precisely because people freaked out that it didn't go into effect.
15:07
So, freaking out is good in Scott Sumner's mind at least."
15:09
Uh anyway, >> [laughter] >> I love this.
15:13
Uh Yeah, we have a solution. >> a solution.
15:16
If if you're feeling the pain at the pump, pivot and get a horse.
15:21
At oil at 110, the urban horse is the only option.
15:25
Pulling up to the gas station on a horse is is truly a lead.
15:28
I I do want to know what's the TCO on a horse with the food and the stables versus, you know, just keeping a What is that?
15:38
A Ford Taurus in your I don't know. That's something else.
15:41
Uh you know, in your garage.
15:44
>> [laughter] >> The horse really mogs at the at the gas station in particular, right?
15:47
Cuz it's just making everyone >> Who would do that? Yeah. Yeah, we got to go back.
15:51
1 horsepower is all you need.
15:53
Um Uh Nick Carter says, "Seeing a lot of non-process trusting panickens on the TL."
16:04
And And then he followed up by saying, "That's what I thought because the oil prices spiked up and then they fell down and we did not get $120 a barrel oil.
16:14
We got exactly 100, which was still >> Still like it's way, way, way too early to celebrate. Yeah. Yeah. Yeah, exactly.
16:22
Uh >> [snorts] >> There's an interesting uh uh chart here, a really shocking chart from Oliver Globe.
16:27
Unreal numbers about Germany's uh nuclear power generation.
16:32
So, if you scroll down, look at this curve.
16:35
This is truly the uh the the bell curve meme or something like that.
16:41
JP Morgan estimates that had Germany not phased out nuclear power, the country would have generated 50% less electricity from fossil fuels and 84% less electricity from natural gas in 2024.
16:53
Electricity prices in Germany would have been around 25% lower and the country would have imported half as much electricity.
17:00
And just complete rise and fall of nuclear power generation in Germany.
17:05
One of the craziest graphs.
17:05
You don't see graphs like that very often in new technologies.
17:09
Typically, you see S-curves or you see exponentials.
17:11
No one considers the The models get better and then they get dumber.
17:19
That's certainly the funniest outcome. Yeah. Um very, very rough.
17:20
Uh Goldman Sachs sent a note to investors saying if oil prices increase by $10 and remain elevated for 3 months, US year-over-year headline CPI inflation would likely rise from 2.
17:33
4% in January to 3% in May.
17:33
Those are small numbers, but we're looking at a oil price increase of maybe $30, $40 over uh the the baseline.
17:42
Uh potentially higher, we don't know where oil's going to land.
17:47
Um and so you have this weird tug-of-war right now with the Fed where uh if if oil prices go up, inflation goes up, the Fed has a mandate to curb inflation.
17:59
That means higher interest rates at a time when the labor market is shedding jobs, you would expect a Fed rate cut or a or a lot of people are are optimistic about a rate cut.
18:09
Trump certainly wants rate cuts.
18:11
Um but if inflation is climbing, there's really no solution other than keeping rates high or or even raising them further.
18:18
So, a real jam in terms of federal monetary policy, Fed monetary policy.
18:23
Uh here's Art Cashin at the opening bell.
18:27
This is a historical video. When was this?
18:31
This was a long time ago. Let's play this clip. This may be it.
18:35
Uh first let me start out morituri te salutamus SA, and you know that's the gladiator salute.
18:40
We who are about to die salute you.
18:42
So, it's going to be a tough morning.
18:46
This may [laughter] be it.
18:48
We who are about to die >> die salute you.
18:52
Insane insane aura for you know, CNBC really doesn't get enough credit for being so innovative in terms of broadcasting and entertainment.
18:59
Uh really some of the some of the greatest clips of >> one of the one of the top comments. This is an old clip.
19:06
Yeah, [laughter] obviously. Obviously, brother.
19:08
480p, it's probably from the '90s.
19:12
Um uh so ZeroHedge says uh you know with oil at 111 total panic and memetic Sisyphus shares a clip says half of the site for the last week is sounded like this.
19:23
Let's >> Producers are telling me there is breaking news.
19:26
The Asian financial markets have just opened to a huge sell-off.
19:32
And we're going to switch to that story right now. Good. I'm glad I'm here.
19:33
Your thoughts Tracy Jordan on how this is going to impact Wall Street.
19:38
Lobby, I'm not an expert, but I do have a strong opinion.
19:42
New York as we know it will no longer exist tomorrow.
19:45
Producers are telling [laughter] me there is Is this from 30 Rock? That's so good.
19:54
>> [laughter] >> That's not funny.
19:55
Oil came way came way in from its overnight highs is Joe Weisenthal and the quote post is Chris Paul >> a huge three to cut down the lead to 42.
20:07
>> [laughter] >> Absolutely absolutely brutal.
20:11
So let me tell you about Gusto, the unified platform for payroll benefits and HR built to evolve with smarter modern with modern small and medium-sized businesses. Stop making fun.
20:22
>> We're we're somewhat of a smarten business. We're small and modern.
20:26
>> Small and modern, smarten.
20:26
Let me also tell you about CrowdStrike.
20:28
Your business is AI, their business is securing it.
20:31
CrowdStrike secures AI and stops breaches.
20:33
So We got to give you this shout-out to George Kurtz for the Mercedes performance yesterday. Oh, yeah. >> One, two. One, two. Really good.
20:40
George Kurtz on Would you look at that?
20:43
Takes a stake in the team, suddenly they're at the top.
20:46
Yeah, he's on to something.
20:50
Well, the my essay this morning was titled why is no one talking about oil?
20:55
Of course everyone is talking about oil.
20:58
Oil discourse gushed onto the timeline this weekend crude prices spiked to nearly $120 a barrel as a broadening war in Iran threatens both transportation routes and production.
21:09
The geopolitical and economic analysis are flowing, but what does this mean for AGI timelines?
21:13
And a lot of people in in the AI world are sort of tuning all of this out because they see recursive self-improvement, AGI, ASI, the build-out as more important.
21:24
And I just wanted to sort of reality check the AI supply chain to understand how does oil actually affect data center construction, AI production, total pricing.
21:34
Like, is there any effect?
21:36
My conclusion was that it's it's very moderate, but there are some interesting effects in the financial markets that are probably the bigger takeaway.
21:44
But, it's still interesting to hear about like, yes, oil actually does is used in the production of AI at least a little bit.
21:51
Um so, power has been at the forefront of the AI pushback.
21:55
Like, everyone's worried about these local energy prices, these electricity prices increasing near the data centers.
22:01
But, and it's been it's been become like a political issue.
22:05
But, pain at the pump might become a bigger story as gasoline prices spike.
22:10
And that has been pain at the pump.
22:10
Oh, it's been the most tangible sign of inflation and moves so quickly.
22:16
You know, one one jitter in the economy >> a huge component.
22:20
You know, people on the coast, people in tech don't have a good sense for this, right?
22:25
If gas gas for a lot of people gas could quadruple and it wouldn't they wouldn't really notice it.
22:31
But, if you actually look into the average American, how gasoline fits into their budget, it's a meaningful component of their monthly budget.
22:40
So, they feel it super intensely. It's variable Yeah. cost.
22:45
>> There were there were so many different ways where a lot of Americans go on driving vacations.
22:48
That obviously is directly impacted by gas prices.
22:50
And then also just psychologically, there's there's something about filling up at the gas tank where you see the number ticking up and you're doing that on a every week basis or so, that it's just so visceral.
23:04
It's this thing that you have to stop and then go experience and watch the money flow out of your account like in real time.
23:10
It's very It's very visual. It's very interactive.
23:14
>> Yeah, I remember I remember I must have been probably 18 at this point where I would just go and I would like for a long time I just put like, you know, prepay for a certain amount of time.
23:26
>> Let's see how much I got.
23:27
>> And I felt like really like I felt like the king of the castle.
23:29
So I just put my card down and let it run up and see how far it go. >> Uh no, no, totally.
23:34
So uh pain at the pump's been like a big uh broad economic issue. It's so visceral.
23:40
Like the actual phrase has a lot of power in politics and in uh like attack ads.
23:45
Like your gas prices are going up under this person.
23:50
That's why you got to vote for me. This is This is common.
23:51
Um but in AI circles, the discussion's been much more focused on RSI now is the new acronym that everyone's focusing on. Not AGI. AGI's here.
24:01
We know artificial general intelligence.
24:03
We passed the Turing test.
24:05
The models can do things generally intelligently.
24:07
Uh but can they recursively self-improve? Are they RSI? Are we in RSI now? Is it a coming?
24:12
Is this going to be a fast takeoff?
24:16
Is this going to be a slow takeoff?
24:17
Well, something's taken off.
24:19
Um Dylan Patel said, "Being in SF is like being in Wuhan right before the pandemic. Something is happening.
24:25
It's going to hit everywhere, but so few people know it."
24:29
So he's sort of echoing that something big is happening mentality.
24:33
This idea of like recursive self-improvement, being AGI pilled.
24:35
Um A lot of a lot of smart people agree.
24:40
There's definitely something happening.
24:41
There's still this question of like, "What is sticky?
24:43
What Where will diffusion take longer?
24:45
What will be sort of AGI resistant or ASI resistant for a variety of reasons, even if they're completely uh irrational and you could get a better thing from a computer, but people like the human version or whatever.
25:01
The irony is that that, uh, George Hotz hitting the timeline to raise money makes me more bullish on acceleration. Right?
25:10
Because if not, he's obviously not a historically been a huge fan of venture venture capital.
25:17
>> not raising from traditional VCs, he said.
25:19
He He specifically said he wants to raise from like, uh, like basically like friends and family type of round. He does want 20 million.
25:26
We have this in the timeline somewhere.
25:28
Where is this post from George Hotz? >> it up if you want. Keep >> Corp.
25:33
He said that he's raising $20 million to buy a building.
25:36
He's going to He's going to buy a building. Here it is.
25:40
If So, he said this is from Tiny Corp, uh, the makers of Tiny Grad and the Tiny Box.
25:45
If Tiny Corp is raising $20 million at $200 million valuation, who'd be interested?
25:50
Business model is basically this. Buy an $11.
25:52
5 million building with 5 MW of power. Link in our Discord.
25:55
Uh, wait for AMD to launch the RDNA 5 96 GB cards mid-2027. Pre-order 3,000 cards.
26:06
Hopefully, we can negotiate for $2,500 each.
26:11
Build 500 $20,000 Tiny Boxes with six of the cards in in each box.
26:16
Run all the Chinese LLMs.
26:19
Make $600,000 per month revenue selling tokens on Open Router market depth is there.
26:24
This is 1% of Open Router.
26:26
Uh, improvements to Tiny Grad yield revenue improvements.
26:28
Due to how power is priced in Oregon, it's only like $50,000 for the electric bill before the 4 MW before, uh, they price for peak, not usage.
26:39
We get like 3 C kilowatt-hour power, 3 cents per kilowatt-hour.
26:45
We can also make $100,000 per month leasing colocation space to comma.
26:50
Building and cards paid off in 3 years max investment made back low risk of being undercut since we're using consumer GPUs and running the cheapest colo you can believe.
27:00
If someone chill wants in, I do it.
27:02
I'm not going to fake hype I'm not going to hype fake tech but demand for tokens is going to skyrocket skyrocket.
27:08
Look at the open claw install numbers with crazy good optimizations.
27:13
We could potentially get 3x more from the machines and we have electricity three for 3x more machines 5.
27:16
4 revenue per month then continue to scale from there custom chips etc.
27:21
He's starting a neo cloud or he's starting a yeah, he's going to be serving tokens.
27:27
Um so that was that was you know incredibly bullish.
27:31
You know demand is definitely there.
27:33
Uh at the same time he he sort of took the other side of like this is different. This is crazy. He says there's no AGI.
27:41
There's no magic threshold.
27:43
You guys see auto research change the random seed from 42 to 137 and OMG it's AGI. It's over.
27:48
Yet you critique the junior engineer for the same stuff.
27:52
The cost of development is falling.
27:55
Overpaid engineering struggles to compete. That's the story.
27:58
And he says humanity has been recursively self-improving for centuries.
28:02
So everything new is old. I don't know.
28:04
I obviously have a huge amount of respect for George but at the same time saying there is no artificial general intelligence when you're also arguing that engineers are struggling to compete with new technology that is replacing them.
28:22
This is how is this not artificial intelligence competing with Yeah.
28:27
So so I think he is he is agreeing with that but it's that it's that there is an immense amount of of desire for this binary moment. This is the singular. This is AGI.
28:39
[clears throat] This is ASI. The RSI is here.
28:40
This thing is happening right now and there's before and after and everything is changed and and and he just doesn't see it that way, I think.
28:49
I think he sees it much more like the internet, the mobile phone, like other technologies that have been rolled out, electricity.
28:54
Yes, there is like a before and after, but you can only really, you know, define the period by maybe a decade, and you need a few decades to understand that moment.
29:03
It's very hard to go back and you know, there is like the iPhone moment, and there is like, you know, the first launch of, I don't know, AOL.
29:12
Like, I don't even know I don't even know what the iPhone moment of the internet was, just because it was sort of a slow rollout.
29:17
Anyway, uh live GPU clusters, um where's Jordy going?
29:24
Oh, you're moving the goal posts. Okay.
29:26
Jordy's moving the goal post.
29:26
Where are you moving the goal post to? Just over here. Okay, but why? Is there a reason?
29:31
Well, I mean, they have to be moved.
29:33
George is moving the goal Oh, okay, okay, okay.
29:34
He's moving the goal posts.
29:39
Uh so, I said, you know, it's clear that the AI industry uh continues to grow and continues to need more and more power and compute, as we've seen from George Hotz's new project.
29:48
Uh that means large data center campuses, uh but if they're not in random office buildings that George is picking up for on the cheap, uh they're probably going to be built with construction equipment. So, what does this mean?
30:01
They they don't just drop from the heavens, uh they require building, which requires oil, but how much oil?
30:06
And is oil a serious in is a serious oil shock enough to impact the AI buildout in a meaningful way?
30:11
Uh spoiler alert, basically no.
30:14
But, um live GPU clusters in the United States do not use much oil directly. Uh only 0.
30:22
6% of US electricity in 2024 came from petroleum.
30:26
We're much, much more dependent on natural gas, something like 42% of US electricity is natural gas.
30:33
And so, uh America ramped up natural gas production significantly over the past two decades.
30:38
A lot of that was in reaction to the the wars in the Middle East.
30:42
Hey, we need to be less dependent on on foreign oil.
30:46
We need to be energy independent.
30:48
And so, you have the fracking boom, the natural gas boom, and that's where a lot of our fossil fuels come from today, I believe.
30:57
Data centers only consume a single digit percent of US electricity. So, you're looking at 0.
31:01
6% of a few percent is like the actual impact.
31:04
So, the short-term impact of high oil prices should be very limited on AI.
31:09
Um building out new capacity is much worse than that.
31:13
>> it's it's worth noting I think I think it's something like Qatar produces something like 20% of the world's liquefied natural gas, right? So, >> But not for us. Yeah, not not for us.
31:24
So, we're in a good spot, but this is going to impact the data center build-outs or or active data centers across the rest of the world if this continues. Yeah.
31:34
So, um >> [cough] [clears throat] >> when you're talking about building new capacity, building new data centers, oil is a little bit more involved.
31:41
So, diesel powers trucks, trains, boats, barges, generators, pumps, compressors, excavators, and tons more >> [music] >> construction equipment.
31:53
Petroleum is also broadly used for plastics, polyurethane, and solvents that all work their way into the data center supply chain.
32:01
The biggest problem is delaying already tight schedules because of narrowly available components going out of stock.
32:06
So, the price of oil goes up, there's one marginal factory that can't produce one ingredient that goes into the rack, and and that slows things down.
32:15
You You have to wait a week while you find another supplier.
32:17
That that stuff can add up to just a little bit of a delay.
32:21
This happened during COVID, and the AI industry was already experiencing something similar with transformers.
32:26
And so, you don't want products getting stuck in transit or going out of stock.
32:29
But the bigger problem, and the one that people should be talking about, and I think you were debating with uh Primack at Axios about this is macroeconomics.
32:36
So, higher prices, higher oil prices lead to higher inflation.
32:40
If the Fed has to raise rates to control inflation, capital formation for mega projects gets a lot harder.
32:44
So, JLL has this estimate, the next 100 GW of data center capacity could require about $870 billion of new debt financing.
32:57
And so, using this rough number, every extra 50 basis points of borrowing cost on $870 billion is 4.
33:02
35 billion in annual interest expense.
33:07
And so, that's a I don't know if that's a huge squeeze on cash flow, but it's certainly a squeeze on cash flow uh if you're a hyperscaler and you're going giga-long AI CapEx now and you're like, "Yeah, we're going to do 200 billion this year.
33:22
We're going to do 500 billion next year or whatever across a couple companies."
33:26
Um you start looking at interest expense bills that are already huge, but then they're going to go up by 5 billion or up by a full more a full extra billion just with a 50% base 50 basis point increase, half a percent.
33:41
So, half a percent uh change in the Fed funds rate can echo through even a quarter percent could probably could probably be amplified into, you know, a a multi-billion dollar uh like cost line on your on your balance sheet or on your income statement um as you're servicing that debt.
33:56
Overall, higher oil prices do make the AI rollout harder, but not really for existing AI capacity.
34:01
I wouldn't expect token pricing to change based on the price of oil.
34:05
Um but there is this more there's this bigger question of macroeconomic resilience during a time where our largest tech companies are digging through the couch cushions to find every penny to win the AI future.
34:14
Uh Middle Eastern investors pulling back, I agree with you.
34:18
I do think that this is a significant risk uh cuz if they pull back on big big investments, makes these mega rounds more difficult.
34:24
Retail investors could fly to oil stocks or defense stocks if they think that this is like going to be a thing for months and months and months.
34:33
It could all flip if there's a quick resolution though, which is what I'm hoping for.
34:36
But if things continue the industry walks an even tighter tightrope.
34:40
So >> mean the the big question right now is the hundreds of billions of dollars of of sovereign AI projects in the Middle East, right?
34:48
I think a lot of those people are going to be like, "Hey, do we want to send, you know, billions of dollars of of GPUs >> Over there?
34:55
And then also the money coming here is another thing where you might want to spend it elsewhere.
34:59
Quickly, let me tell you about TurnRout Puffer, serverless vector and full-text search, both from First Principles and Object Storage, fast, 10x cheaper, and extremely scalable.
35:06
And let me also tell you about Public, investing for those who take it seriously, stocks, options, bonds, crypto, treasuries, and more with great customer service, whatever your investing strategy is.
35:16
>> happy birthday to Trey. Happy birthday, Trey.
35:19
>> very excited to have you. Thank you.
35:19
One of our strongest soldiers. We appreciate you.
35:23
It's been an honor podcast with you.
35:26
>> noticing that it's been hotter in Los Angeles? I have.
35:29
>> Downtown Los Angeles is forecasted to approach a 100° F on Thursday and Friday in March.
35:33
Which is why >> Not good.
35:38
We're going to do the weather.
35:40
New segment for you today.
35:40
We're doing the weather on TBPN.
35:43
We have our very own Ben. We have Ben. Hello guys, how are you? We're doing great.
35:48
Tell us about the weather. What's happening?
35:50
Well, I want to start off by saying, as you can see, the weather today for the low low today is going to be 75° F up there, high of 100° F down there.
35:59
But there's something I actually wanted to point out that I saw and I thought that was quite interesting.
36:03
As you can see up here, there's a localized low pressure area up there and a localized high pressure area right down there if you can see that right there. >> mean rain?
36:10
The issue with that's not normally an issue and not a cause for concern, and it's not very common for this time of the year.
36:15
However, today later in the afternoon these two areas are going to collide.
36:18
They're going to hit each other. Really?
36:20
What that's going to cause is a barometric pressure inversion.
36:24
It might sound a little bit scary, but I guarantee there's no cause for concern.
36:28
All that means is that hot air rushing in from the west is going to collide with that cold air rushing in from the east and it's going to cause a bunch of turbulence in the sky moving all the air waves around and oscillations in the sky.
36:38
However, I want to add one more point.
36:41
A byproduct of this effect is that all that humidity that dropped after that hot air moved to the bottom is going to raise up cuz the water cycle cycle you know evaporation and stuff it's going to raise up into the sky into those clouds going to cause big clouds in the sky.
36:53
And eventually all that water is going to fall down onto the ground.
36:56
We're going to have [laughter] big rain later in the afternoon.
37:00
Wait, is it is actually going to rain in LA?
37:02
Yeah, it might sound crazy, but I just wanted it for all you guys at home.
37:06
I'd definitely try to step outside with a jacket today maybe a hoodie just in case the rain comes.
37:09
Don't try to be a big shot be prepared.
37:12
>> fact checking you right now and the weather app says it's going to be sunny all week.
37:15
Is this just complete fake news?
37:18
>> [laughter] >> My team those are the numbers my team gave me.
37:20
I don't know I can't >> fake news. No, no, no, no.
37:21
This is the fakest news I've ever heard. >> no.
37:24
The all that the transatlantic current John, John, I'm I don't trust your app. I trust the weather app.
37:31
Apple says it's not going to rain the entire month.
37:36
Have you looked at the transatlantic Did you look at the transatlantic current? No.
37:39
My precipitation 0 in today, 0 in tomorrow, 0 in on Wednesday.
37:45
>> John, you're really going to you're really going to trust an You're really going to trust an application that was probably vibe coded yesterday over Ben who's doing the weather. >> he get here? What happened here? >> [laughter] >> No.
37:59
If you look right there you can see a localized high pressure area.
38:00
That's a localized high pressure area.
38:04
>> weather man who's just looking for drama in the most boring weather market in America which is Los Angeles.
38:10
>> I think I think this is the most important story Look at this this is ridiculous.
38:15
>> can look at the jet streams.
38:15
They're coming in from the west.
38:16
I don't want to hear any more mumbo jumbo about jet streams. Get out of here. You're done. work, Ben. you, Ben.
38:24
>> [applause] >> The fake news weather will return, maybe never, who knows.
38:27
Let me tell you about Restream.
38:29
One live stream, 30 plus destinations.
38:31
If you want to multistream, go to restream. com.
38:35
And let me also tell you about Gemini 3. 1 Pro.
38:37
With a more capable baseline, it's great for super complex tasks like visualizing difficult concepts, synthesizing data into a single view, >> future.
38:46
>> creative projects to life.
38:47
>> This is the future of the weather.
38:47
You have a weather man Yes.
38:48
getting into a live altercation with one of the other hosts >> who's just constantly fact-checking.
38:55
>> It's on site next time, Ben. It's on site.
38:57
If you pull up with some fake news on TVP and weather report, it's on site. >> Good job.
39:03
So, it is It's not Adam Smith's birthday.
39:08
It's the 250th anniversary of the publication 250th Sorry.
39:10
250th anniversary of the publication of The Wealth of Nations. Tyler, you have a copy?
39:19
I I have two of our copies here.
39:19
I didn't bring So, yeah, usually I have them with me, but I didn't bring mine to work because we have a we have a few here.
39:26
>> Yeah, yeah, yeah, yeah, obviously.
39:28
For those who don't know, give us the pitch for Wealth of Nations.
39:31
Why should people read Wealth of Nations in 2026?
39:33
Yeah, I mean, this is kind of the OG, uh you know, economics book, right?
39:37
Was it It was base Like, put it into like AI terms.
39:41
This is just like situational awareness type of or or not type Attention is all you need.
39:47
>> like the ChatGPT moment for capitalism. Yes. Yes.
39:50
But I I would like to you know, I have a few quotes I would like to read. >> Okay.
39:54
Uh all right, let me open that. Let's see. It's a big book. It's like 800 pages.
40:00
And it's really, really dry.
40:02
I mean, it reads like it was written 250 years ago.
40:05
Okay, so this is I think this is kind of one one of my favorite quotes.
40:08
Kind of the, you know, seminal.
40:11
Every individual necessarily labors to render the annual revenue of the society as great as he can. Mhm.
40:14
He generally, indeed, neither intends to promote public interest >> [applause] >> nor knows how much he he is promoting it. Yes.
40:24
He intends only his own gain. Yes.
40:24
And he is in this, as in many other cases, led by an invisible hand to promote an end which was no part of his intention. Yeah.
40:34
Rational economic actors, homo economicus, thinking of the the the the value for society is created by rational self-interest.
40:44
And it's sort of the spin on like greed is good.
40:46
It was like he was like an OG of that of that thesis. >> maxer.
40:50
Uh yeah, productive labor, efficient markets, uh free markets, don't hold gold, don't control trade.
40:55
An important read in 2026. Somewhat heretical.
41:01
There are many people that disregard that book. Maybe they should not.
41:04
Maybe they should give it a crack at it. >> Evergreen. I think it's evergreen. I think it's evergreen.
41:07
I mean, there was that debate over like Tyler Cowen over the weekend.
41:11
And somebody was taking shots at uh Stubborn Attachments for being like not like not advancing the the discipline of economics.
41:19
And my pushback to that was that I actually think it is an incredibly it's just a fun read.
41:26
It's like really well written and it's very entertaining and it's very thoughtful.
41:29
But it's particularly interesting to hear someone sort of just rehash the tenants of liberal democratic capitalism in 2020 I think he wrote it in 2023, 2022, something like that.
41:43
But like like there there there are a lot of voices out there who are anti-capitalist, anti-democratic, anti-liberal, and Tyler Cowen is just retreating to like a time-tested, time-honored uh economic philosophy that is deliberately not revolutionary.
42:05
And I thought that it was a great book for that reason and that's why I recommend uh Stubborn Stubborn Attachments to so many people and so widely.
42:12
Um but uh if you want to crack open the real the real the real hard stuff, pick up a copy of The Wealth of Nations.
42:21
Why is one of those copies so much smaller?
42:22
Is the font size just smaller?
42:25
Uh yeah, it's super small. It's also two columns. >> Oh, two columns. Wow.
42:28
Yeah, if you can get through that, you're you're truly you're truly elite in the global economy.
42:33
Uh what did the timeline have to say about uh the the the masterwork that is turning 250?
42:41
It's now needed more than ever, says the Washington Post.
42:43
On March 9th, 1776, 4 months before the American colonies broke with Britain over the issue of taxation, a little-known Scottish thinker published a long, dense book with an unpromising title. Shots fired.
43:00
An inquiry into the nature and causes of the wealth of nations.
43:02
250 years later, the uh Adam Smith is by any objective measure easily the most widely cited and widely quoted economist who ever lived.
43:12
Astonishingly Astonishingly, his work still frames the central questions we face, not just about free markets, trade, and capitalism, but about the nature of human society itself and even what it is to be human at all.
43:25
Smith was born in 1723 in Kirkcaldy, Scotland.
43:30
Educated at the Universities of Glasgow and Oxford, he initially made his name not as an economist, but as a moral philosopher.
43:39
His first published book, The Theory of Moral Sentiments, offered a radical theory of how we form moral judgments.
43:44
Radical because it derived from the creation of moral values not from scripture or divine grace, but from human sympathy and mutual regard.
43:49
The Wealth of Nations, as his second work, second major work came to be known, was an extension of that product.
43:56
The book is not, as sometimes believed, a hymn to greed, a paean to market fundamentalism, or red in tooth and claw capitalism.
44:04
It was an attempt to understand how a commercial society could generate prosperity without collapsing into corruption.
44:11
This 25th This 250th anniversary is not a moment for hagiography.
44:16
It is an opportunity to recover a way of thinking that is directly relevant, indeed urgent, to the economic, social, and political challenges we face today. Begin with trade.
44:26
In his own time, Smith's great target was mercantilism, or what he called the mercantile system, the idea that wealth consists of hoarded bullion, and that trade is a zero-sum contest.
44:36
Governments granted monopolies, imposed tariffs, and manipulated commerce in the purported pursuit of national strength.
44:42
Producers were widely protected, consumers often ignored altogether.
44:46
Against this, Smith >> days, we should have Tyler just do a speed reading of The Wealth of Nations live on the show.
44:52
As long as you can >> Speed run?
44:56
Uh yeah, Wealth of Nations speed run, any percent, glitchless.
44:58
You need to do glitchless, though. >> Glitchless.
45:02
Oh, for sure, glitchless run is the only way we can do it. No glitching. Uh no no AI summaries.
45:06
I think that would take you probably 5 days. Probably 5 days. It's so dense.
45:16
>> Have you seen how fast he can rip a Rubik's Cube?
45:18
>> This edition is 488 pages around.
45:18
Oh yeah, I guess you don't have to read all the footnotes and stuff.
45:22
You can just kind of skip through it, but it is a it is a thick book.
45:26
But the But the way it's written is just really not fun to read.
45:30
It's like it it reads like it's 250 years old.
45:32
It's not It's not like a page-turner.
45:35
Well, we need to talk about the wealth, but first, Vibe. co.
45:42
Where D2C brands, B2B startups, and AI companies advertise on streaming TV.
45:45
Pick channels, target audiences, and measure sales just like on Meta.
45:46
And let's also tell you about Fin.
45:48
ai, the number one AI agent for customer service.
45:51
If you want AI to handle your customer support, go to Fin. ai.
45:53
We need to talk about the wealth of Sundar Pichai. Pichai.
45:59
Sundar >> Sundar new pay deal worth up to 692 million.
46:06
Is this like 10 times what Tim Cook's making?
46:08
No, it must be over over time, right?
46:10
>> And Cook's making around 70 a year. >> 70 a year. >> Combined.
46:15
>> So if he works for 10 years, he makes what Sundar makes in three?
46:19
There's going to be a conversation.
46:22
>> I mean, but this is what we've been we've been We've been advocating for this. So this is good. >> Yeah, no, no.
46:25
We're we're we're we're we're in support.
46:29
>> has increased Sundar's potential pay to 692 million over the next three years.
46:35
>> members group chat as soon as it hit.
46:35
He was just like, "Dude, this is this is a cool article. You should read this." He just drops it in. Just drops it. Thoughts? Yeah, exactly.
46:44
The bulk of his package comes in performance units with a target value of 126 million split evenly into two tranches.
46:53
The PSUs are valued by the parent company, blah blah blah.
46:54
It could pay out as much as twice the target or a quarter billion if it outperforms significantly or nothing if it lags behind. >> beat the S&P 100.
47:05
Um Pichai will receive Waymo stock with a target stock in Waymo with a target value of 130 million and 45 million in Wing Aviation.
47:13
That's their um uh their their drone delivery >> platform. platform.
47:20
Uh again, both can pay out up to 200% of the target.
47:26
>> He's got a lot of incentive pay.
47:26
Like if he if he delivers in the and the company does well, um he should he should be richly rewarded.
47:33
Uh Barron's was writing about the the death of the Mag 7. Mag 7's over.
47:41
Uh Up and Down Wall Street, the Mag 7 era ends. Barron's is calling it.
47:43
They said, "Stick a fork in it. Turn out the lights. Hasta la vista.
47:49
Say it any way you'd like.
47:53
The simple truth is that the Mag 7 trade is over. Finito."
47:54
I love uh Barron's writing. Dead.
47:57
The collective stock market outperformance of those seven tech icons, Alphabet, amazon.
48:02
com, Apple, Meta Platforms, Nvidia, Microsoft, and Tesla is now a thing of the past.
48:09
The group may still do okay, and some of the individual stocks may even kill it, but the slam dunk set it and forget it run circles around the market era of the Mag 7 is gone with the wind.
48:20
Recall that BofA Securities analyst Michael Hartnett coined the term Mag 7 in 2023, pretty recently.
48:25
Um referring to the 1970s John Sturges Western gunslinger flick, an adaptation of Akira Kurosawa's 1954 film Seven Samurai.
48:38
Have you seen Seven Samurai? No.
48:41
Have you seen The Magnificent Seven? The stocks?
48:47
>> [laughter] >> No, the movie, the movie, the Western. No.
48:49
He's done the Seven Samurai. In what world?
48:52
Production team, has anyone seen either of those movies?
48:55
I think we watched them in school. >> Okay. Okay. Oh, you did.
48:59
>> Which one did you watch?
48:59
You're saying you didn't watch them, you studied them.
49:00
That's what you're saying.
49:02
I haven't I haven't seen them in a while.
49:03
>> Which ones have you seen, Tyler? Seven Samurai.
49:05
Seven Samurai, it's a great one.
49:06
It's a It's a very It's a very good movie.
49:08
Anyway, uh Barron's is bearish on the Mag 7, but Sundar Pichai is doubling down.
49:14
>> It does sound like cope.
49:14
I think the Mag 7 will do >> It's like the large It's over for the largest and most profitable companies in history that stand to benefit >> Yeah. some ways.
49:23
argument was the growth to value narrative, the loss of cash flow as they increasingly invest in CapEx.
49:30
That the the financials will look very different.
49:32
There'll be margin compression, that type of thing.
49:33
Uh it it it's not an unreasonable take, but it is is just funny the way it's written.
49:39
Anyway, Sundar Pichai took over as CEO in August of 2015.
49:42
He's going on 11 years in the seat.
49:45
Google's market cap has increased almost sevenfold from half a trillion to 3.
49:51
6 trillion briefly topping 4 trillion in January.
49:53
This surge has made the Indian-born 53-year-old former McKinsey consultant a billionaire.
50:00
He joined in 2004 and made his name developing the Chrome browser and leading the Android division.
50:05
He had been criticized for being too slow to adopt AI at the search giant, allowing Open AI to release the first first hit product ChatGPT in late 2022, but has since bounced back, releasing cutting-edge AI models and integrating the technology into its dominant search engine.
50:22
Yes, he's done very, very well on that front.
50:23
Pichai has also navigated a duo of antitrust cases brought against Google's search and app store businesses, avoiding the worst-case scenario of forced breakup.
50:32
A third lawsuit is pending against the advertising network.
50:36
Tyler, do you think that DeepMind counts as as RSI capable at this point?
50:44
Recursive self-improvement.
50:44
So, there was a take on the timeline where basically someone was making the argument that xAI is cocked because well, Anthropic is using Claude to build Claude and OpenAI is using Codex to build Codex.
51:00
So, both of those leading private AI labs are recursively self-improving.
51:07
xAI was using Claude code to build Grok and so they don't have the same flywheel of the tools that build the tools build the tools.
51:15
>> Yeah, cuz I think they lost access to Claude code, but then they still have access to Codex. Okay.
51:18
So, you can it's kind of like do you count that as RSI cuz it's not like really internal? Yeah.
51:24
Um But is Gemini that's the question that Yeah, yeah, it did it did it did it did it did it did it did >> people at Gemini are using Gemini.
51:28
I mean there is a Gemini CLI.
51:30
People don't talk about it as much. Yeah, yeah, yeah.
51:34
I mean the model is very good, right? >> Yeah.
51:35
Yeah, I I It would be interesting.
51:37
I feel like we don't get as many little data points from the Gemini and DeepMind team because it's a public company and they need to, you know, maybe break out financials at some point.
51:48
>> Yeah, I think like vague posting is is probably, you know, more heavily looked at. >> Exactly, exactly.
51:54
Um, because you would yeah, yeah, because it's totally possible that they're doing the same thing.
51:58
I mean, Demis has been extremely, uh, just very clear about the path this goes and he's very AGI pilled.
52:04
He understands the exponentials and whatnot.
52:05
So, uh, he's certainly messaged it at a high level without actually saying, you know, definitively, yes, like we are at a point where we're not writing any code by hand.
52:16
That that that type of phrasing that you see in the in the RSI crowd.
52:20
Uh, anyway, Pichai got last got a stock award in December of 2022 worth 218 million, which was structured in the same way.
52:28
His earnings are topped up by his personal security costs, which rose to 8. 3 million in 2024.
52:32
Uh, earlier in the week, he sold 32,500 Class C shares at an average price of $303 worth roughly $10 million.
52:40
The Bloomberg Billionaires Index estimates that he has sold about 650 million dollars in stock since becoming CEO.
52:49
Um, they still own, uh, along with his wife, he owns 1. 7 1.
52:54
67 million shares of Google worth half a billion at the latest stock price.
52:59
And, uh, Google's founders, Sergey Brin and Larry Page, still control the company through their ownership of super voting Class B stock, which gives them 56% of decision-making power. Uh, interesting.
53:12
I think you got to give Miami >> paper hands. Oh.
53:16
>> [laughter] >> He sold 650 million of stock. >> No, no, no.
53:19
He sold that, but he's, uh, oh, and he only owns half a billion.
53:22
So, yeah, he sold more than that.
53:26
>> hopefully hopefully hopefully he put it all into video. >> Paper hands Pichai.
53:32
But, he's getting topped up.
53:33
And he's diamond hands the new he he's diamond hands the new CEO.
53:36
>> just saying I'm just saying, looking at the stock chart since 2015 when he became CEO, he would have done pretty well just to not do anything and trust trust his own process.
53:45
>> Hey, hey, maybe he has an amazing car collection or something, you know?
53:48
There's many There's a lot of things Get to rotate into metal, man.
53:50
>> You can't drive a Google class class A share.
53:56
>> [laughter] >> You can't You can't sleep on a bed of class C shares, Jordi.
53:59
You got to You got to get liquid.
54:02
Um No, you got to give Miami tech its due because everyone's saying, "Oh, Miami tech, they don't have any startups.
54:09
There's no funding there.
54:09
There's no Miami tech isn't panning out."
54:11
Well, Google is now counts as a Miami tech company because the Google founders bought houses there.
54:18
You also got to count Meta as a Miami tech company because Mark Zuckerberg has a house there.
54:23
So, you put Meta and Google, that's like $6 trillion in market cap for tech companies in Miami. I count it. I count it. Anyway. Moving on.
54:32
>> Uh owner of Four Loko is exploring a sale of a of the storied alcohol brand, sources say.
54:37
Well, let's dig into this, but first, let me tell you about Plaid.
54:42
Plaid powers the apps you use to spend, save, borrow, and invest securely connecting bank accounts to move money, fight fraud, and improve lending now with AI.
54:48
And let me also tell you about Cognition.
54:49
They're the makers of Devin, the AI software engineer.
54:51
Crush your backlog with your personal AI engineering team.
54:54
So, The parent company of Four Loko, the canned alcoholic beverage that became a college campus sensation in the late 2000s before being reformulated under regulatory pressure, is exploring I cannot believe they nerfed Four Loko.
55:08
>> [laughter] >> It's so >> It was too powerful.
55:12
Like you you can just imagine the trajectory of the United States if it hadn't been nerfed.
55:18
And then >> Straight downward. Really? >> Yes. It was so bad.
55:20
It was so the impact >> people were having like heart attacks. Yes, it was terrible.
55:25
So, Four Loko was a 24-oz can, so twice the size of a normal alcoholic beverage.
55:32
>> Trey, the birthday boy, says it was amazing. >> Okay.
55:34
Now we have BuzzBallz, that's true.
55:36
Um So, Four Loko, the original formulation was twice as big as a normal can of Bud Light or Miller Lite or Coors Light, your normal beer, something that you would grab.
55:48
Um so, it was it was it was two drinks that way.
55:51
And then instead of it being uh 3. 2% alcohol or 4.
55:54
5% alcohol, it was like 10% alcohol.
55:58
It was the strength of like wine almost.
56:00
And so, a single Four Loko was like four or five beers in one can.
56:08
And then they also added like 200 mg of caffeine.
56:11
So, you were you would you would become incredibly intoxicated and inebriated, but then also incredibly stimulated from all the caffeine.
56:17
And that spelled doom for many people.
56:22
People would be very high-functioning, but completely um inebriated and and discombobulated.
56:26
And so, they would uh you know, get into all sorts of trouble whether or not it was at the behind a motor vehicle or >> Discombobulator.
56:34
It was the It was the It was the discombobulator.
56:36
Uh it was also wildly illegal from an FDA perspective.
56:40
You can't mix alcohol and caffeine in a single product.
56:42
That's just a rule that they chose not to follow.
56:47
And they figured out how to make it.
56:49
And then uh the FDA eventually gave them a warning letter and they pulled the caffeine out and that of course killed the viral sensation that was Four Loko, but apparently the company continued to sell products that were compliant with the current FDA regulation. Nerfed. That's nerfed.
57:08
That's the nerfed Four Loko continued to sell to sell to >> they're looking to uh and potentially will capture somewhere around 400 million on the sale. >> value.
57:20
So, people It's hilarious that they're working with JP Morgan on this. That's wild.
57:23
So, Fusion Products is the owner of the brand.
57:26
They're working with JP Morgan on the sale process sale process.
57:29
Uh the potential sale underscores how ready-to-drink beverages have emerged as a growth category in an otherwise sluggish alcohol market.
57:37
US beer and wine US beer, wine, and spirits sales declined in 2025.
57:43
We've talked to a number of guests about the declining alcohol sales as folks move over to more functional beverages. >> CMO had to say.
57:50
For over a decade, our sales have been the leading flavored malted beverage have have been leading the flavored malted beverage market by embracing bold innovation, unconventional marketing, and a risk-taking attitude that delivers results year after year.
58:08
I guess they kept building.
58:10
They also own >> The parent company also owns Mamitas, Pirate Water, >> Pirate Water, >> Boscobel Tequila, and just Earthquake. Earthquake?
58:20
Earthquake seems like something you definitely want nothing to do with. >> Loko, Earthquake.
58:25
>> [laughter] >> Pirate Water is a wild name for a brand, too.
58:30
I mean, it's a distinctive can. It looks remarkable.
58:32
It jumps off the shelves.
58:33
When these when the Four Loko Nerf went into effect and they pulled them from the shelves because before they reformulated without the caffeine, they had to pull them from the shelves entirely.
58:42
They were a hot commodity and folks would go around stockpiling them before they were banned and and they were the one of the most guarded resources in all of college campuses.
58:54
>> Well, Earthquake is just a can that's just 10% alcohol. Wow. Do not recommend. Very risky. Never quake.
59:01
Anyway, in 2010, the FDA sent warning letters to caffeinated alcohol beverage makers including Four Loko saying that the caffeine in their products constituted an unsafe food additive.
59:11
So, the FDA has specific categories for regulation around around what is an alcoholic beverage, what is a caffeinated beverage, and they never the two shall mix.
59:21
So, they don't let you put these two things together.
59:25
Of course, vodka Red Bull was the precursor to this where folks would order a Redbull and then they would put vodka in it, I think, but at a bar, which is made at the time of sale and so that was legal, I believe.
59:39
And this might be fake news, but I remember hearing a story that at one point Redbull was sued because a gentleman had purchased a series of vodka Redbulls, become incredibly inebriated, gotten behind the wheel of the car of a car, gotten into some sort of accident and wound up in front of a judge.
59:59
And the argument was that he wanted Redbull to be held liable for his actions while inebriated and behind the motor vehicle.
1:00:11
And his argument was that it was only because of the product Redbull had been mixed with the vodka that he made the decision that he was alert enough, stimulated enough from the caffeine to get behind the car wheel in the first place.
1:00:25
And this might be apocryphal, but the story goes that Redbull made the argument that the caffeine actually made him more capable to drive, not less capable to drive, and it actually increased the safety of the of the driver and the vehicle, and so they were held not responsible.
1:00:41
But I don't know, it's just like an old wives' tale in CPG.
1:00:44
I don't know if it's true.
1:00:46
Anyway, let's move on to Phantom Cash.
1:00:49
Fund your wallet without exchanges or middlemen and spend with the Phantom card.
1:00:52
And let me also tell you about Labelbox.
1:00:55
RL environments, voice robotics, evals, and expert human data.
1:00:58
Labelbox is the data factory behind the world's leading AI teams.
1:01:02
So, we're all preparing for the singularity.
1:01:05
What is this doing dishes?
1:01:07
Just like, oh, if you're if you're doing dishes, you're not you're not worried about Dylan Patel's, you know, viral post saying that being in SF is like being in Wuhan right before the pandemic.
1:01:18
Something is happening, it's going to hit everywhere, but so few people know it.
1:01:20
And the that is the question of of are we in a software only singularity where the the dishwasher is life as normal um in and and every time I go to the COVID analogy, I start a clock.
1:01:32
Because I remember tracking COVID in January, and I remember the viral post virology posted in February going viral.
1:01:40
And then in March, the NBA shut down, and and in April, everyone was locked down.
1:01:46
And so, every time I see one of these, it's like COVID. I'm like, "Okay.
1:01:51
Let's see what it Let Let's see Let's see what it's like in 8 weeks.
1:01:53
Like 8 weeks is not that long.
1:01:54
In 8 weeks, if if I can if I go outside and things look radically different, and you talk to random people, and it's wildly different, then we're like, "Okay, yes, that analogy was correct."
1:02:05
But if it's more like in 8 weeks, we're like, "Oh, yeah, like you know, technology's getting more exciting, and like tech companies are changing, and like the market's moving, but like, you know, the the the the the the the dishwasher is unaffected, the the the pizza place on the corner is unaffected."
1:02:20
That That's always how I I I I just think people underrate like how crazy a global plague was.
1:02:25
Like, it's literally like a once in a hundred like the most insane thing ever.
1:02:29
It's like the the craziest change to literally everything from the sports that you watch.
1:02:35
You couldn't go to the movie theater, you couldn't go to dinner.
1:02:38
Like, everything was changed for like a full year.
1:02:40
It was like complete upheaval in our society.
1:02:42
And might happen, but 8 weeks 8 weeks is a little fast for me.
1:02:49
And if you live in LA, probably still be able to go to the beach. I think so.
1:02:51
So, that's But maybe not.
1:02:54
Maybe everything's Maybe maybe the Uh I did have a funny moment yesterday I was telling you, but I'll I'll share it again.
1:02:58
Uh I was talking with my my uh car detailer who's become a dear friend, and uh he was telling me that uh the people in my area Mhm.
1:03:08
uh cancel their appointments at a much higher rate than uh the people in the rest of LA. >> Mhm.
1:03:15
And he was talking about running analytics on iMessage, and I was like [clears throat] how how are you how do you do how do you do that and he's like he's fully running open claw.
1:03:23
So he's now running using open claw to run his like car detailing business.
1:03:29
He's like yeah just just always running in the background.
1:03:33
It's it's some to him he was like it's some combination of having like an EA >> Yep.
1:03:38
or a CEO like helping me manage a calendar helping with comms helping me make smarter decisions on routing and scheduling and all these things.
1:03:46
So when you book a detailing meaning that this gentleman is coming to your house and cleaning your car you send an iMessage to him correct? >> Yeah.
1:03:58
And that is what is so interesting about open claw is that there were probably a thousand companies that were like we're vertical SAS for auto detailing.
1:04:07
Uh you got to do everything over email.
1:04:10
And they just got eaten alive because people use iMessage and iMessage has been this closed ecosystem that >> Yeah.
1:04:15
private companies could never access before.
1:04:17
But for but it makes so much sense for him to say hey run through all my iMessages and if I'm talking to my friends about the the the the the F1 race just discard that but if I'm talking to a client about potentially booking a detailing put that in a spreadsheet organize that and then run some analytics and tell me what the patterns are.
1:04:37
And that's just something that should have been SAS for a decade >> Yeah.
1:04:43
but Apple has a walled garden that has now been smashed down by open claw and I think it's really cool.
1:04:49
>> It and I think that I think that that's what we're seeing the first like use cases for open claw.
1:04:52
It's like things that could not happen for business reasons for logical reasons sometimes for legal reasons if you're talking about like downloading illegal films like the Napster analogy but a lot of times it's just it's just Apple didn't want to do a deal to integrate iMessage into Salesforce.
1:05:10
>> The way to break down the walled garden was just to have a digital guy. Basically, yeah.
1:05:13
Who just Who just combs over all this.
1:05:15
So, uh is that a new type of software that that should have existed that now is unlocked? Is it disruptive?
1:05:22
Like, it's not like he It's not like he stopped using his previous CRM.
1:05:26
He probably didn't have anything.
1:05:27
So, now he has something.
1:05:29
Yeah, he he's pretty tech forward.
1:05:29
He's tried every single SaaS tool that there is.
1:05:35
But, like, at the end of the day, having like a digital guy that has access to all of your tools is just a much better experience.
1:05:41
And so, I want to know what his bill is.
1:05:45
>> [laughter] >> Uh this month.
1:05:46
>> He said he spent a decent amount on I want to know what it is this month.
1:05:49
And then, I want to know what it is in, like, a year.
1:05:53
Because sometimes the maintenance of these systems can be really really big.
1:05:56
And then, there's a question about are we going to see per token deflation?
1:05:58
Or are we going to see prices Are prices being artificially held low because of uh competitive dynamics?
1:06:04
So, uh in a year, will he be spending more or less?
1:06:08
Will he still be getting the value?
1:06:09
These are all interesting questions, but I'm just excited that a a digital product now exists that could not or didn't exist before.
1:06:16
And that feels incredibly positive and cool. So, very exciting.
1:06:21
Anyway, we have our first guest of the show in the Restream waiting room.
1:06:24
Let me tell [music] you about console. com.
1:06:28
Console builds AI agents that automate 70% of IT, HR, and finance support, giving employees instant resolution [music] for access requests and password resets.
1:06:35
And without further ado, Dr.
1:06:37
Alex [music] Wiskott Gross, welcome to the show. How are you doing? What's happening? Doing really well. Thank you, John. Thank you, Jordy.
1:06:42
Thank you so much for taking the time to come chat with us.
1:06:45
Uh congratulations on all the virality.
1:06:47
For those who uh haven't followed along so far, can you introduce yourself and the project broadly?
1:06:52
And then, I have a ton of questions, but I'll let you kick it off with an intro. Absolutely.
1:06:57
So, quick background on on myself.
1:06:59
Originally from New York, undergrad MIT, studied physics, electrical engineering, computer science, math, PhD physics Harvard.
1:07:04
Was there anything didn't study?
1:07:06
You [laughter] studied everything.
1:07:10
Didn't study the humanities. Okay.
1:07:12
Considered Consider going to Cambridge to study chemistry and biology for a while. Just started it.
1:07:15
Was told physics is a game for the young. Okay.
1:07:18
>> So, stayed in the US, did my PhD in physics focusing on AI and nanotech. Cool.
1:07:22
I've started, invested in, advised, managed dozens of companies at this point.
1:07:26
I've had a number of exits.
1:07:28
And at this point, most of my time is focused on smoothing out the singularity, as as you guys were just discussing. Yeah.
1:07:34
So, most recently, as as you were gesturing, a company that I helped found, Eon Systems Public Benefit Corporation, just announced what we've been characterizing as the first multi-behavior upload, if you will.
1:07:48
More on that in a second.
1:07:50
Of a fruit fly, which we think is a major step forward for the field.
1:07:55
And and quite frankly, in an era when you know, I I do another podcast, the Moonshots podcast, with my friend Peter.
1:08:02
I I talk all the time about tiling the earth with computer.
1:08:04
Right now, all of this AI infra build-out that we're doing to the tunes of trillions of dollars of CapEx, all of this is going to artificial minds.
1:08:12
And Eon is playing, I think, a very important role in doing some early pioneering experiments and developments to try to level the playing field of the future.
1:08:22
Not just artificial minds playing on all of this trillions of dollars of infra and data centers, but enabling emulations, and if you will, uploads of human minds, non-human animal minds, at the moment starting with fruit flies and aspirationally working towards mice and humans, and getting getting the brains of a variety of human and non-human animals to operate in the cloud. Mhm.
1:08:49
So, walk us through the fruit fly experiment.
1:08:51
Why it was so exciting for you? It's really interesting.
1:08:55
If you think back, remember back with the launch of ChatGPT, that was just a side experiment.
1:09:02
That that wasn't the main focus of OpenAI at the time.
1:09:03
Some have characterized that as an un-hobbling, if you will.
1:09:08
GPT-3, the the leadership of OpenAI thought at the time, was sort of the the main entree, if if you will.
1:09:16
Similarly within Eon, I don't mind saying this was more of a side project.
1:09:20
That the we thought it would be somewhat interesting to to the world, but I I think the the level of response and virality probably even took most of Eon by surprise.
1:09:30
So, this this is a video demonstration of an internal project to take a bunch of building blocks that were already sitting around for the past 1 to 2 years and put them together for the first time.
1:09:42
So, one of these was a paper that senior scientist at Eon, Philip Schou, published in Nature in 2024, demonstrating that just from an emulation perspective, that you could take the FlyWire, which was well-funded project to capture the connectome of the fruit fly brain, that you could emulate certain circuits within the fruit fly brain fruit fruit fly brain.
1:10:06
We've had that for a year and a half or so.
1:10:11
Another project, NeuroMechFly version 2, is an excellent project to to be able to formulate simulations in realistic mechanical environments of fruit flies.
1:10:22
Another project from Ersland et al.
1:10:22
to formulate coordinated motor actuation of fruit flies. And then AI.
1:10:30
So, in in summary, Eon was really the first to put all these building blocks together that were sort of lying around.
1:10:37
This is actually tiny subset of what Eon's working on.
1:10:40
Eon is working on scanning the connectomes of of human and non-human animals at scale and achieving whole brain emulation, which is the the holy grail of the field.
1:10:51
There have been many projects in the past that have attempted human whole brain emulation.
1:10:56
Eon is leveraging advances in AI, expansion microscopy, a variety of other techniques, but it was really just putting the pieces together.
1:11:04
So, we're quite quite honestly astonished by the the level of interest in just this one demo. I love it.
1:11:08
I have a million questions.
1:11:10
Let's start with uh uh IO.
1:11:13
Like the inputs and outputs.
1:11:13
Like what are you simulating?
1:11:17
When I pull up The Sims, that is somewhat of a simulation of a human.
1:11:21
The you know, a a basic for loop over a bunch of random probabilities to say, "Are you hungry?"
1:11:29
If hungry, eat, go over to the stove, cook food.
1:11:32
Uh that's obviously nowhere near the level of detail at which a human operates.
1:11:36
Um is there an abstraction layer here where the output is quantized to things that you would think a fly would do, like move a uh like move a wing or it or are you operating at the level of like choose to fly or how at what level of abstraction is the fly's output happening?
1:12:01
Yeah, this is what I was gesturing at a moment ago.
1:12:03
So, there's a paper by Erdel et al.
1:12:06
that bundles motor actuations in terms of high-level representations.
1:12:10
So, a lot of people sort of the the people who got perhaps most excited in the neuroscience community about this video, some of them I think didn't bother to read Erdel's excellent paper.
1:12:22
So, a lot of those This is why I'm saying a lot of the pieces were sort of lying on the ground waiting [snorts] for us to pick up and put together.
1:12:29
So, the these like these abstract motor movements were already available.
1:12:33
Just no one had ever prior to this bothered to wire them up. That's cool.
1:12:39
Um then what is the Do you need a system prompt?
1:12:43
Like do you need like an like an initialization function?
1:12:45
Or once you once you uh synthesize the connectome, synthesize the connections in the brain, uh uh things just start firing.
1:12:54
Do you have to trigger something to to start the process of consciousness or whatever you call whatever the fly is doing? No system prompt needed.
1:13:04
On the other hand, in in the early days of ChatGPT before instruction tuning, there was no system prompt either.
1:13:10
In the days of early language models, no one had thought of prompt engineering.
1:13:15
So, it's entirely possible that that maybe in the future some sort of prompt engineering equivalent for for whole brain emulation will be essential, but at the moment there there's no direct analog of that.
1:13:30
Help me compare some of the scales that we're talking about on complexity of a fly, complexity of a mouse, complexity of a human.
1:13:38
I imagine that we're on some sort of exponential here, but uh how many orders of magnitude in complexity, size, number of neurons, something like that, whatever the the metric is, how far away are we from humanity?
1:13:54
We're many orders of magnitude away.
1:13:54
So, the the canonical estimate for the human brain is depending on how you count brain cells, and there are multiple types of brain cells, call it order of magnitude 100 to 200 billion cells in the brain.
1:14:07
We're many many orders of magnitude Many orders of many orders of magnitude away from that.
1:14:14
So, this isn't going to happen immediately.
1:14:16
We're not going to get human whole brain emulation >> Yeah. immediately.
1:14:19
On the other hand, I I think it's important I think back to the early days of driverless cars and autonomous vehicles.
1:14:26
And what the world is missing right now is I think sort of a scale or a framework for starting to think about what that future looks like.
1:14:34
We're at the earliest stages of that.
1:14:36
Call it level one or level two uploading.
1:14:39
A lot of people even get triggered by the term uploading.
1:14:40
They prefer something else.
1:14:42
And I think we need to start having these conversations in a critical thoughtful way and start to define what is a framework for multiple levels of fidelity for uploading or emulation or maybe we come up with a new term, but right now we're starting that discussion and we're trying to do it in a thoughtful responsible way.
1:15:00
So you like the term uploading because you did not uh you did not choose a random structure or an average structure of this particular fly's brain, you chose the exact structure of a particular fly's brain.
1:15:15
This is a very expensive structure of a very particular fly's fly's brain.
1:15:18
This was the the the the structure was again captured through the the flywire project at great expense. Mhm.
1:15:27
Uh so that's like a clone digital twin type of thing, right?
1:15:31
That is that >> Aspirationally.
1:15:32
I mean again, we're at the earliest stages.
1:15:34
I I don't want to oversell this is just a prototype and and quite frankly, this is a sliver of everything that Eon is doing, but the fact that Eon was able to put a few of these pieces together, build a prototype with relatively small effort suggests that, you know, we talk in the AI community all the time about overhangs.
1:15:54
Many people thought that the arrival of large language models was in some sense a technological overhang that we could have one could imagine a thought experiment.
1:16:02
Could we have had large language models 20 years ago if we simply knew what we should have been building? I think Could we have?
1:16:09
I thought we were very GPU gated on that.
1:16:13
Well, the GPUs were being spent on video games for for the first few decades.
1:16:15
So we we could have in principle like Markov models, we we knew how to build Markov babblers.
1:16:21
I was building Markov babblers in the '90s and early 2000s.
1:16:25
Yeah, it just feels like even if you even if you try and port back frontier models now to a you know, 1080 Ti which was a graphics card for gaming, uh you're going to have a bad time.
1:16:39
Well, I mean we're with the rise of SLMs and some of the the amazing deflationary properties of algorithmic progress in in SLMs, I think we will find ourselves in the world next 2 years where some of the earliest PCs could probably have hosted non-trivial conversations with LLMs.
1:16:58
There's just that much algorithmic progress.
1:16:59
Extremely bullish for Apple.
1:17:02
>> [laughter] >> Anyway, not to put it in in hardware terms. What do you want to do?
1:17:05
What do you want to do next?
1:17:06
Like where do you Where do you go from here?
1:17:10
I know it's a I know it's a side question.
1:17:11
More flies or straight to mouse?
1:17:15
It It's such a profoundly interesting question.
1:17:17
So, Eon is beginning fundraising.
1:17:19
Eon wants to to tackle both mice and men. Ooh.
1:17:22
And I I think the time scale will be determined by both technical capability increases and also fundraising abilities.
1:17:32
So, I I think it's an interesting time.
1:17:34
I I I like to say uh on my other pod that the singularity looks like all sci-fi tropes happening everywhere all at once.
1:17:42
And one of these one of these sci-fi tropes is most definitely mind uploading.
1:17:47
And I'm doing my best vis-a-vis Eon systems to push that part a little bit to the left so that the AIs don't have the Dyson swarm all to themselves. Interesting.
1:18:00
Uh architecturally, what does the neural network look like that you've built for the fly?
1:18:07
Is it Is it similar to a large language model >> like a transformer-based architecture or are you a beneficiary of that?
1:18:15
To- To- To- totally unlike the transformer.
1:18:16
Tran- Totally unlike the transformer.
1:18:18
So, transformer architecture, if you subtract off the encoder and decoder layers, looks and comes in many variants, but in the the most vanilla variant looks like alternating linear and attention layers.
1:18:32
So, totally unlike the transformer architecture, at least if you look at it.
1:18:35
It looks like more like a a little bit more like a graph neural network, but really that that's such an AI way of framing it.
1:18:41
It looks like a leaky integrate fire LIF model of that that we've had for for decades from the neuroscience world.
1:18:48
So, it's it's just a graph of nodes that have leaky integrate fire dynamics and they're firing at each other.
1:18:55
So, what does that tell you about the nature of Sim synthetic human intelligence?
1:19:02
Like do you think that there is a path to human level AGI that does not involve the transformer and involve instead falls along your path and and your architecture? I'm sure.
1:19:15
Well, first of all, I I would argue we've had AGI for at least 6 years now, 5 to 6 years at at at the very latest since since summer since summer of 2020 when Open AI published their language models or few shot learners paper, yeah, GPT-3.
1:19:28
I would argue that was an AGI.
1:19:30
So, I I I think there are many ways.
1:19:32
My bet is there are many ways to achieve generality of intelligence.
1:19:35
I published a paper a number of years ago arguing that not only is intelligent behavior a general process, I I went further and argued that it is a general physical process that you can even formulate intelligent behavior in pure thermodynamic terms.
1:19:50
So, my bet is [clears throat] intelligence is this very very general effect and lots of ways to implement it. Interesting.
1:19:57
Do you So, do you I mean, you've mentioned AGI 6 years ago, singularity here and you're smoothing it.
1:20:06
Do you have other sort of binary benchmarks that you're looking towards like ASI, recursive self-improvement?
1:20:14
Are any of these terms useful to you? Oh, we're already there.
1:20:17
I mean, we're we're already in the era of recursive self-improvement.
1:20:20
All of the frontier labs are are pretty public about it.
1:20:22
At this point, I'm looking past the singularity.
1:20:26
So, I spend most of my time on bets for what the post-singularity world looks like.
1:20:31
So, Eon is one of those bets.
1:20:33
Another bet, I have a company Physical Superintelligence that's trying to solve all of physics with AI and doing an amazing job.
1:20:39
I just wrote a book with Peter called Solve Everything arguing that entire disciplines are going to get steamrolled by superintelligence and that what matters now is what disciplines we aim the superintelligence at. Interesting.
1:20:54
What What decisions do you make in your personal life in the way that you live >> Yeah.
1:20:59
that are based around your beliefs around uh technological progress?
1:21:07
There are a few that come to mind.
1:21:07
So, one is I'm hoping not to die.
1:21:08
I It would be a a shame if if I died sometime soon and get to miss the most exciting developments. Same.
1:21:19
That's what gets the applause? That's great. Yeah. That's a white pill. White pill, very good.
1:21:24
Bryan Johnson thanks me in advance.
1:21:28
Uh so, uh but I I think also, I mean, there there are so many other angles and another is not trying to bet against the collective intelligence of the market.
1:21:39
I've had a a number of friends who who thought they could outsmart all of the AI algo traders and day trade and they they when they inevitably fail, they attribute it to bad luck or or something other than than just fundamentally betting against progress and betting against AI intelligence.
1:21:57
So, trying not to make the mistake of betting against collective intelligence of civilization and AI capabilities.
1:22:05
I I also I'm I'm trying to make bets assuming that the AI capabilities keep increasing to the point where they're ontologically shocking and trying not to duplicate effort of frontier labs.
1:22:15
And I I think that's perhaps something I don't see enough of in the venture community.
1:22:23
So many new startups being formed now just aren't being ambitious enough.
1:22:25
I have startups that newly formed or recently formed that are literally trying to grow new islands and new coastlines with AI.
1:22:34
That would have been unthinkable years ago.
1:22:36
I I have >> What is that that's that's with physical AI?
1:22:43
Or or they're just dreaming them up?
1:22:46
Uh it's with AI reaching into the physical world to to steer ocean currents and grow new islands and new coastlines.
1:22:52
That would not have been possible a few years ago.
1:22:54
That's now starting to become possible.
1:22:56
I have a startup that's working on solving interspecies communication starting with dogs. Love that.
1:23:01
That would not have been possible a few years ago.
1:23:03
So I I spend a lot of my time thinking about what the post-singularity state of the world looks like and how to smooth that out and bring it here sooner.
1:23:12
Yeah, what does the post-singularity world look like where you are able to upload yourself?
1:23:16
I assume that feels like there's a copy of Alex in the cloud and then you still exist in not in the cloud.
1:23:23
But how do you think that that interfaces?
1:23:29
In in my mind if if we wind up having the copy question Yeah.
1:23:31
come about, that's almost a failure mode.
1:23:33
I I speaking for myself, I don't want a copy of myself. I I want myself. >> Yeah.
1:23:40
And Hans Moravec and others have written about this.
1:23:42
What I would like to see in the space and what Eon is working toward at least aspirationally is this idea of a continuous transfer of consciousness.
1:23:51
So that it really is you. It's not a copy of you.
1:23:53
It's not a low-fidelity facsimile of you.
1:23:55
It should be a better expanded version of you that's still you.
1:24:00
And whether that looks like moving from what Eon's doing right now Yeah.
1:24:05
with fruit flies to maybe replacing a single cell in a brain at a time invasively or non-invasively with a substrate-independent or substrate-migrated implementation, some variant of that I suspect is where all of this will go so that you never have to worry is it really my self or not.
1:24:21
It will by construction be yourself.
1:24:23
And I'm I'm not sure of the precise timetable, but I I I think 5 10 years from now, I think the world will see marked progress on the the problem of not just whole brain emulation, but also transfer of human intelligence to new substrates. That's very exciting.
1:24:45
This is Very new sci-fi corner. I love it.
1:24:47
Uh well, congratulations >> all at once. Time travel? >> good framing.
1:24:53
>> Is time travel on the table? We don't know yet. Okay.
1:24:56
>> So, it's an interesting question.
1:24:56
We don't know whether the physics of our universe are compatible with I think what most folks would construe as as time travel.
1:25:03
There are There are versions of temporal non-locality that are consistent with the physics that we have right now, but not in a useful way that you would that would be worth filming a sci-fi movie over.
1:25:13
What about uh faster-than-light travel? We don't know yet. It's same problem.
1:25:18
Um in in some sense, exactly the same problem as time travel.
1:25:22
We don't know whether the physics of our universe will ultimately allow that.
1:25:26
>> Cuz it seems like if you don't get [clears throat] FTL, like there's a it's going to be a boring travel to go to another solar system.
1:25:34
I I will say this, I think odds are pretty good that AI in the next few years will tell us whether the the laws of physics of our universe are compatible with that or not and help us solve this.
1:25:42
One of the reasons why I helped found Physical Superintelligence to to discover any of such laws. Very cool.
1:25:49
Well, thank you so much for taking the time to come chat with us.
1:25:52
>> Yeah, great to meet you.
1:25:53
>> Congratulations on the progress. Fascinating stuff. >> Thank you. Likewise. We'll talk to you soon. Cheers. Have a good one.
1:25:58
Let me tell you about Figma.
1:25:58
No matter where your idea starts, Figma make, quad code, codex, or a sketch, the Figma canvas is where ideas connect and products take shape.
1:26:07
Build in the right direction with Figma.
1:26:09
And let me also tell you about Cisco.
1:26:13
Critical infrastructure for the AI era.
1:26:15
Unlock seamless real-time experiences and new value with Cisco.
1:26:17
And [clears throat] without further ado, we have Charles Lamanna from Microsoft.
1:26:25
In the research ready room.
1:26:25
Now he's in the TV room from Charles. How are you doing? Good.
1:26:29
Thanks for having me on, guys.
1:26:31
Thanks so much for coming on the show.
1:26:33
Uh why don't you give us a little bit of an introduction on yourself and then we can get into the news. Sure thing.
1:26:38
I'm on the product and engineering side at Microsoft.
1:26:40
Uh I my team builds out things like agents, a bunch of the the platform work on our business applications and and some low-code stuff.
1:26:48
So a bunch of odds and ends.
1:26:49
And what uh what's the latest and greatest in your world?
1:26:53
Yes, a lot of big news today.
1:26:53
So today we announced Copilot Co-work, uh which really takes Copilot beyond chat.
1:26:58
So it can take actions and automate tasks and pick delegation, basically.
1:27:03
We had some new agentic capabilities in the Office apps and a bunch of agent management solutions.
1:27:08
Yeah, how are you thinking about uh the the open claw boom, people writing code locally versus Microsoft has Azure.
1:27:18
Like it's so easy to throw everything in the cloud and I feel like uh one of the barriers that a lot of people have is like, do I need to go get a Mac mini?
1:27:27
Can I just delegate to the cloud?
1:27:27
Some of the apps don't just by default build an app in the cloud or they don't have functionality.
1:27:34
How are you feeling about that trade-off these days?
1:27:38
Yeah, I mean I we're super excited by open claw and all that means.
1:27:41
I think it's like the resurgence of the PC.
1:27:42
Like what's the best thing you can give an agent? A computer, you know? Let it run the apps. Let it log in as you. Let it message as you.
1:27:49
Uh I think like a Windows PC in the cloud is probably the main way people will ultimately use something like open claw at work.
1:27:58
Yeah, how do you think about uh model selection?
1:28:00
Obviously, Microsoft has a has a time-honored partnership with OpenAI at this point.
1:28:05
But uh you know, Microsoft has their own AI lab and uh you can get pretty much any model on Azure.
1:28:14
Do customers want to pick models?
1:28:14
How do you think about what is what's the best tool for the job in the products that you're building, but then also how much do you want to give to the user cuz there's sometimes when I go into a chat app and I'm like, I just want auto.
1:28:27
You should be figuring this out.
1:28:29
I I know the details of pro versus thinking, but I don't want to think at all.
1:28:34
Yeah, so that's one of the things that we think is super important for Copilot and a lot of our products is the fact that it's multimodal.
1:28:40
And the reason is because the best model for a given kind of task is not always the same and it's not even the same from month to month or quarter to quarter.
1:28:48
Like if I look at Microsoft 365 Copilot, chat is like the main way people interact with it.
1:28:53
Our auto router is going to put you in open AI models cuz that's what we think works best with our work grounding and our work context.
1:29:01
But then if you go to Co-work, the new announcement today, and you fire off a task and you want to run for a long period of time in the background, use a bunch of tools, maybe generate some code, use like the terminal and stuff, that uses the Claude family of models.
1:29:13
And >> [clears throat] >> I go to Copilot one place and get both and then we're going to keep adding more models, models like from MAI and other places where it makes sense, but it's really about the experience the user wants, not so much picking exactly the version or the type. Yeah.
1:29:29
How have conversations been with business leaders in terms of deploying a product like Copilot Co-work?
1:29:32
Because the ramp up for software engineers was smooth and everyone was aware of it, tracking everything, and going from, you know, better auto complete to agentic coding was was pretty intelligible at every step of the way for most software developers, but for someone who's lived their life in spreadsheets and email and maybe not really thought about building automations all that often, what does actual deployment of these tools look like?
1:30:07
Yeah, it's such a great question cuz I think there's like two pieces to it.
1:30:09
The first is you have IT teams which are nervous with if you imagine every single employee spinning up a whole bunch of agents that can create documents, send emails, run workflows.
1:30:23
Um so they want to make sure they have the right governance and controls in place.
1:30:26
That's why another big part of the announcement today was something we call Agent 365 which makes it easy for IT teams to manage and deploy agents from Microsoft and other companies. So that's one part.
1:30:35
And then you have the end users.
1:30:37
How do you teach a typical Excel or PowerPoint or Outlook user this idea of agentic working?
1:30:43
I mean even in coders it it's taken time and it's probably not fully diffuse.
1:30:47
Um so that's the thing that we think we're in a really unique position at Microsoft in terms of being able to do cuz we can bring things like co-work to users where they already spend their time inside of the Office apps, inside of Windows, and increasingly inside the Copilot app.
1:31:03
Uh and we can present it in a way where it's digestible to hundreds of millions of Office users, not just tens of millions of coders.
1:31:12
Is the majority of your business model still seat-based at this point because I imagine that there will be a transition at some point.
1:31:19
How much of that transition's already happening?
1:31:23
Yeah, we're we're still very seat-based on like the M365 side but like the good news at Microsoft, we kind of know how to do consumption businesses cuz we got Azure, we got a bunch of other very large capacity businesses.
1:31:33
So the fact that we have user licensing and then consumption and pay-as-you-go licensing for capacity all integrate with one commercial kind of vehicle with customers, we think we're in a pretty good position to make that transition over the next couple of years.
1:31:47
There was some uh there was some chatter on X this weekend over uh how different labs are pricing tokens and plans potentially to take market share from each other.
1:32:00
There's a little bit of a capital fight playing out.
1:32:02
How are you communicating, or more specifically, competing with the application layer? >> Yes, yeah.
1:32:07
And so, and so, my question is, how are you talking to business leaders, CTOs about how they should think about forecasting token budgets and spend and compute budgets if they if they're used to, "Okay, I have licenses and licenses maybe go up in price and my head count goes up and I'm used to that."
1:32:27
But all of a sudden it's like, "Well, everyone's paying $200 a month, but it might be $5,000 a month in a couple months, but also the models are getting cheaper and the inference is getting cheaper and the capabilities are going up, and it's very hard to predict."
1:32:38
What are you talking to What are you telling to business leaders and CTOs?
1:32:44
Yeah, so I the the first thing is you have to embrace this technology, otherwise you're going to be left behind right now because the productivity gains are just so significant.
1:32:51
And if you go talk to a coder and you told them, "Hey, I'm going to take away like GitHub Copilot and the agentic coding capabilities," they'd be like, "I refuse to work in this environment."
1:33:01
Like it's just it's just inhumane almost.
1:33:03
So So I think that the same type of thing is going to happen for all information work, all office work.
1:33:10
So 9 months from now, I think if you went to somebody and said, "We're going to take away your agentic tools like Copilot Co-work," they'd be like, "No way, I'm not going to go back to the old way of working."
1:33:20
So So I think the first is there's a degree of inevitability because the benefit is so large and there's such strong pull from the end users.
1:33:30
That said, you got everybody has budgets.
1:33:32
So what you're going to see and what you're already starting to see is a bunch of controls where IT teams can start to manage exactly how many tokens or how much dollars each user can spend and they're going to allocate that in a way based on where they think they're getting return and diminishing return and it'll probably be like a expense report approval hierarchy type thing.
1:33:51
So you'll see all of that show up, but part of just doing work now is you need a token budget.
1:33:59
That that's the future of how we think about and how you budget for a team or a department. >> Mhm.
1:34:05
Where do you have any exciting case studies on the non-software engineering agentic work side of the business?
1:34:16
Uh because when when agentic coding came to software development, uh you know, there was a shift in a lot of software developers uh workflows where they went from writing a lot of code to reading a lot of code, reviewing PRs.
1:34:27
They They started operating at higher and higher levels of abstraction, but the decision what to build, where how to prioritize things, how to build features, there's still conversations and software engineers are still doing work.
1:34:40
Uh what does that look like in the context of uh of a finance department or an operations department or a sales group?
1:34:47
Like where when the rubber meets the road, like where are you seeing like, "Okay, this This is ready.
1:34:53
This is the next uh this is the next sub area or category that will really really benefit from the productivity gains from agentic workflows."
1:35:04
Yeah, I mean, a general thing and then maybe a more specific thing.
1:35:06
Uh what we're seeing is like you have How do you measure productivity? Is itself a challenge?
1:35:11
I mean, this is this predates AI, but right?
1:35:13
Like lines of code, number of pull requests, number of bugs fixed.
1:35:18
The I don't think I've ever heard a good engineering leader say that's how I measure the success of my team, right?
1:35:24
It shows up in other ways.
1:35:24
Do you have a great product people love?
1:35:26
Is your cost structure good?
1:35:28
What's your install base look like?
1:35:30
So So I think we have to get better at capturing that final business outcome cuz that's what's going to get the CFO to open up the pocketbook and continue to invest uh in more tokens and more AI.
1:35:42
So if I go then take that and apply to kind of other disciplines, like you you What you don't want to say is, "Oh, our sales people send a lot more emails."
1:35:50
Or our finance professionals create a hundred more spreadsheets.
1:35:53
Like that's more like of a nightmare scenario for [laughter] AI.
1:35:57
So One billion spreadsheets per company. >> Yeah. No, no.
1:36:00
I mean yeah, we saw someone on the timeline say that their their token budget went way up, but their revenue didn't go up.
1:36:05
And so they're like, I don't know if it's uh if it's worth it.
1:36:09
Like you got to show results. Exactly.
1:36:12
So that that is what the question's going to be everywhere.
1:36:13
But the the good the good news is um engineers historically have resisted clear measurement of impact cuz you know there's a little bit of art to it.
1:36:21
But like sales professionals or finance professionals or customer service and and support reps, they are so metric driven.
1:36:31
Like if I'm a sales rep, what's my quota? What's my attainment?
1:36:35
What's my revenue per head?
1:36:38
Every single sales department has all of those things instrumented.
1:36:40
And we've seen with M365 Co-pilot, we kind of did a bunch of experiments and we published a white paper on it.
1:36:46
Like an almost 10% increase in revenue per seller just by giving them this the family of M365 Co-pilot tools and things like our sales extension.
1:36:55
Like that's real dollars in the bank.
1:36:59
Same thing for like back office finance processes.
1:37:01
We've seen people use like our agents to do like accounts reconciliation.
1:37:05
And we have like a great case study we did there where it was like 30% of all back office finance was automated out by this thing.
1:37:12
So those those are examples of real top line and bottom line benefit.
1:37:16
And that that's what everybody's going to be talking about the next couple of years.
1:37:19
It's not going to be, you know, emails, documents, files, etc. Yeah. Yeah.
1:37:25
Uh I mean you're in a you're in a high-level business role yourself.
1:37:26
I find myself doing a lot of like knowledge retrieval, deep research reports.
1:37:33
But have you personally been using agentic uh you mean this product specifically in your day-to-day?
1:37:38
Like even as something as simple as like a Sam Altman told us that he has a self I I he has some sort of like to-do list that does itself or something.
1:37:46
I still don't understand exactly what that product is, but how are you using these tools these days?
1:37:53
Yeah, so the for for like Copilot teamwork, for example, I I've I've loved that.
1:37:58
Like my favorite magic moment has been around calendar management.
1:38:00
I was I was telling the story to someone like I go go look [clears throat] out 3 months in the future and try to see, you know, how crazy is my calendar and travel look like. Interesting.
1:38:10
>> And I was able to decline 17 meetings for me.
1:38:12
Okay, that's my favorite right there. I don't know why.
1:38:14
Nothing gives me more joy than declining a meeting. You sold me. Yes.
1:38:19
AI And I'd say like I went to Outlook and just saw all the meetings disappear and it was like such a serene moment for me.
1:38:25
And And like it did a deep analysis.
1:38:27
It went and said like who's my manager, who do I work with, am I optional or required, what's the topic?
1:38:33
And it recommended the 17 and I went through all 17 and said, "You know what? Decline." And it was right.
1:38:37
I mean, so like that's one thing that I use it for.
1:38:40
Um like another thing that I like right now is I mean, I'm an engineering background, so I'm not going to go create the most beautiful content all the time.
1:38:48
So I use it now to pull together information, design docs, meeting notes, and then go create either like the architecture diagrams, presentations, or spreadsheets, or docs. Yeah.
1:38:58
And I mean, I know it's just more doc generation, but it's very much more kind of like fire and forget. So works really well. Yeah, yeah.
1:39:04
I mean, even if it's just like the bones of a of a PowerPoint presentation with the actual data pre-populated that would just previously be a lot of copy paste, go run this analysis, copy it over, make sure it's the same font.
1:39:17
Like there's so much grunt work that can be automated with these tools.
1:39:21
I'm super excited for that.
1:39:23
Um I I W- W- Where do you see the the design PowerPoint generation uh tool use?
1:39:29
I mean, people have been talking about like it's hard to create an RL environment with a verifiable reward around a great PDF, but it's easy to do that with coding.
1:39:38
Do you Do you have like an internal data set of like the best PowerPoints ever created and you're that you're like are relying against or where do you see that going?
1:39:49
Yeah, I mean I think that is something that there's a little bit of taste in all of this cuz like depending on the company, some companies like lots of infographics and lots of colors. Yep.
1:39:59
Other companies like super fact-based, no images.
1:40:00
So like you know, there's a little bit of an eye of the beholder, but absolutely we have a a very robust harness and like gym environment where we're able to quickly evaluate models, prompts, tools, and our own like fine-tuning efforts to see if it's producing better content.
1:40:15
So I think like this like a PowerPoints, Word documents, spreadsheets, it's going to be like code. Yeah.
1:40:22
And you're starting to see it already, but it's going to be like code for sure.
1:40:27
Yeah, I mean like it a lot of these things are like XML under the hood anyway and and can be puppeteered like code anyway.
1:40:33
Well, congratulations on the launch.
1:40:35
Jordan >> you don't mind I'm calling co-pilot co-worker Coco. Coco.
1:40:41
So, hopefully hopefully it doesn't stick unless you like it.
1:40:47
>> Yeah, it's a very cute name.
1:40:50
Uh but congrats congrats on the launch and great to meet you.
1:40:52
Thank you so much for taking the time to come on the show. Cheers, Jordan. We'll talk to you soon.
1:40:56
Let me tell you about MongoDB.
1:40:56
What's the only thing faster than the AI market?
1:40:59
Your business on MongoDB.
1:40:59
Don't just build AI, own the data platform that powers it.
1:41:04
And let me also tell you about Graphite code review for the age of AI.
1:41:08
Graphite helps teams on GitHub ship higher quality software faster.
1:41:10
Uh there's an interesting data point from Apollo featured by Restructuring here.
1:41:17
Uh AI is making starting a business easier and the numbers confirm it. Did you see this?
1:41:22
Weekly business formation is exploding higher.
1:41:25
So, uh previously >> this matches what the Collison brothers said when they were on the show.
1:41:30
They're seeing not just more business formation, but businesses growing Yeah. faster.
1:41:34
Is this it could be that AI is actually making businesses harder to start, but but Stripe Atlas is making it so much easier that we're seeing an upward trend, and it's it's [clears throat] fighting back against that.
1:41:47
Now, of course, these are >> seniors are just so out there though.
1:41:51
>> are multiplicative, very clearly.
1:41:51
Uh but Stripe Atlas does make it very easy to start a business.
1:41:55
Uh US business applications jumped insanely after the COVID pandemic.
1:42:00
Uh in 2019, we were seeing about uh 60,000, 70,000 new business applications.
1:42:08
Uh I think a four-week moving average.
1:42:09
Uh we've been over 120k for years now, and AI seems to be driving that trend upwards by about 20% over the last 2 years, which is uh great news.
1:42:22
More More small businesses, they are extremely edifying, extremely valuable. I highly recommend it.
1:42:28
>> got another white pill. Give it to me.
1:42:30
The data center build-out has allowed Loudoun County to lower property taxes by 38% since 2010, translating to average savings of about $3,400 per year for homeowners. This is great. Love it. What's happening here?
1:42:45
Data center square footage in the millions has gone from 5 million square feet to 40 million square feet. That's a huge jump.
1:42:55
>> Yes, this is like Northern Virginia.
1:42:57
This is like AWS East one. >> Okay. >> Yeah.
1:42:59
So, so on homeowners are paying they went from 1.
1:43:02
2% of whatever their house is.
1:43:05
So, if they have a million-dollar house, they will be paying $12,000 a year, and now they're at 0.
1:43:13
8% property tax rate cuz they're taxing those data centers.
1:43:16
This is the easiest solution to every AI problem.
1:43:19
If the companies are profitable, tax them.
1:43:21
If the If we already tax every business, we we tax them by default.
1:43:26
If somebody makes a lot of money on an AI stock, tax it as capital gains. You get the money. The money shall flow.
1:43:33
And so, good news for the folks over in Loudoun County.
1:43:36
Jacob says America once named things plainly, place, trade, surname.
1:43:41
You had business, International Business Machines, Long Distance Telephone. Studebaker. Studebaker. And Bethlehem Steel.
1:43:54
>> That International Business Machines logo is goes unbelievably hard. Is this real?
1:44:00
Or is this like a reimagining of what the IBM logo could have been because I love this logo. What a very cool thing.
1:44:10
Um That is apparently the original IBM logo. >> logo. Wow.
1:44:13
We don't know how to make logos like that in this country anymore. Seriously.
1:44:17
I mean, that that's what kicked all this off.
1:44:19
Somebody was was upset about the granola rebrand America's innovation slowdown is primarily due to the amount of Lord of the Rings that founders have to flip through before identifying a name they can raise off of.
1:44:28
No, you can just start a company and name it International Business Machines.
1:44:32
Whatever you want to do, just name the company exactly that.
1:44:36
Technology and Business Programming Network. It's that easy. In other news, 2 Chainz.
1:44:42
2 Chainz revealed he was an early investor in SpaceX. This is huge.
1:44:44
He says "It just felt like I got in like early Bitcoin."
1:44:50
>> some crazy early SpaceX investors.
1:44:50
I don't know what Elon was doing with that early road show.
1:44:54
We know some of these folks.
1:44:56
Very cool community early SpaceX >> Get me 2 Chainz.
1:45:00
2 Chainz got in at some point.
1:45:03
And Gabriel over at OpenAI says 2 Chainz is about to be 3 Chainz.
1:45:06
Well, let me tell you about Okta.
1:45:11
Okta helps you assign every AI agent a trusted identity.
1:45:14
So, you get the power of AI without the risk.
1:45:16
Secure every agent, secure any agent.
1:45:18
Let me also tell you about the New York Stock Exchange.
1:45:21
Hopefully, we'll be seeing SpaceX there soon.
1:45:23
Want to change the world?
1:45:25
Raise capital at the New York Stock Exchange.
1:45:27
Can you imagine two chains and Elon ringing the bell at the New York Stock Exchange? I can't. I hope it happens.
1:45:33
Well, we have our next guest in the Re-stream room, [music] Julian Back from Sequoia Capital. He's a partner there.
1:45:39
Welcome to the show, Julian. Good to meet you. >> Hey guys. How you doing? Nice to see you.
1:45:45
We also hope we can ring the bell very soon. Holy holy logo mark. >> Yeah. Holy logo mark.
1:45:51
>> either of those companies? I do.
1:45:51
We're we're we're partnered with Cisco.
1:45:53
Thank you for backing them early, making TVPN possible. 30 years ago Cisco.
1:45:57
Uh Yeah, Cisco paved the way.
1:46:00
Since this is the first time on the show, why don't you give us just a brief back story on yourself, how you wound up at Sequoia, what you focus on, what your day-to-day is like? Sure.
1:46:10
Um so first that thank you both for having me.
1:46:12
Uh it's uh it's an honor to to be there.
1:46:15
I know you had a lot of my partners as well, so we're very excited to to be on the show. Yeah. Um my name is Julian.
1:46:21
I'm a partner on the team. I'm based in London. Mhm.
1:46:23
But I spend a lot of time going back and forth uh to spend time with the rest of the team in California.
1:46:29
And I started as a founder very very long time ago in high school. Wow.
1:46:34
Uh built companies in hardware before. Go.
1:46:36
Uh and so that's why I've been investing mostly in software um for most of my career. Yeah.
1:46:43
Well, you you you were like this stuff I got to get out of here.
1:46:47
Are you thinking about going back to hardware? Is software dead? What are you thinking?
1:46:54
Yeah, well, I mean two things can be true at once.
1:46:56
I think hardware is really interesting. Mhm.
1:46:58
Um I'm spending a lot of time thinking about physical AI. Sure.
1:47:02
Uh but at the same time I've been investing for the last 10 years in application software. Okay.
1:47:07
Uh so I want to remain true to that uh and this is kind of the things I've been spending time on recently.
1:47:13
Yeah, take me through your current thesis on how software is changing in the age of AI, in the age of you know uh you know maybe some sort of takeoff.
1:47:24
AGI is here according to Sequoia Capital amongst many other people.
1:47:31
The labs are recursively self-improving and it seems like you know Sam Altman made the the claim that like if you want to build a startup, you shouldn't build a company that isn't direct in the direct path.
1:47:44
Well, more and more products seem to be in the path of AI and AGI.
1:47:46
How are you thinking about the changing dynamic with software as we know it?
1:47:54
Yeah, that's a really important question and I think we've been listening to Sam a lot as a shareholders.
1:47:58
Um and recently just talking to our founders, a lot of them were wondering if they're just an iteration away from the models replacing what they're doing. Yeah.
1:48:10
And it puts us in the position of thing what is coming next.
1:48:15
What what is going to be defensible? Yeah.
1:48:16
Um and for us we we've been making that prediction last week that the next trillion-dollar company will be a software business that masquerades as a services firm.
1:48:30
And we shared that thesis with with everyone online.
1:48:34
The the reality is if you sell uh tools today, you're really in the line of sight for the models and you're effectively competing with the next generation that they're they're going to launch.
1:48:46
Whereas if you sell the work, you're actually benefiting from what the models are doing and all the billions of dollars that are going towards AI.
1:48:55
Uh so part of that prediction was saying for every dollar that's spent on software, $6 are spent on services. Sure.
1:49:03
Um until now we could only really go after the $1 because we were building tools.
1:49:08
Now that we can actually deliver outcomes, companies should think about capturing those $6 instead.
1:49:17
So, the I mean the obvious like first step is just get to outcome-based pricing, move away from a seat-based model, probably.
1:49:25
Would you agree with that as sort of like the most immediate change in business structure that needs to happen?
1:49:32
Or if you're starting a new company, maybe don't start with seat-based pricing on day one?
1:49:37
Or is the like seat-based pricing is dead, maybe maybe that's like an overblown characterization?
1:49:43
How would you How do you think about just like the business model moving forward?
1:49:49
Yeah, we we think both can work.
1:49:49
They just serve different purposes. >> Okay.
1:49:53
Um we have companies in our portfolio that started as these co-pilots, right?
1:49:58
That are helping the workers.
1:49:59
And then we call them auto-pilots, the one who, you know, do everything end-to-end.
1:50:03
Many of them are transitioning from co-pilot to auto-pilot to to do what you're describing.
1:50:08
A company that comes to mind is a business called Sierra that was started by a Bret Taylor. Yeah.
1:50:14
Um that basically helps companies with customer support, which is pretty direct bull's-eye use case of AI. Yeah.
1:50:22
And in this example, they're actually charging the customers for outcome, right? For resolved ticket.
1:50:30
Uh and so if you spend $20 per ticket resolved with humans, they charge you, say, $5 instead, right?
1:50:36
And And that's driving ROI for the company, it's driving an outcome that's very measurable.
1:50:41
And so that's the first more obvious company that's really working doing that.
1:50:47
Uh frankly, until recently, outside of software engineering, we haven't seen a lot of this sort of, you know, fully owned uh end-to-end outcome driving.
1:50:55
Um but the models are really progressing quickly.
1:51:00
And for everyone who's been bite-coding the last couple of months, uh I think we've all been been impressed that we can code a full website just end-to-end without any skills, frankly. Yeah.
1:51:12
Um and and basically our prediction is that this is going to happen to other categories.
1:51:15
Um and we've been mapping out these categories and helping founders come up with frameworks to prioritize where to attack them and how to do it. Yeah. Sorry to interrupt.
1:51:27
One thing I'd I'd be curious to get your take on.
1:51:29
So, uh if a company is just selling uh enterprise software that helps people accomplish a task that feels very threatened by, you know, uh foundation model company that maybe makes an agent that can also do said said task and then you don't even need the software.
1:51:47
So, if a company transitions to just selling uh selling the work, selling the service, like you're saying, uh you would agree that they're still going to be under under threat from the labs who will also presumably be selling intelligence or agents that can go and do the work, but you're saying it will maybe be more closer to a fair fight.
1:52:05
Is that Is that the right way to think of it?
1:52:09
Because I still think uh let's let's say if somebody is like spinning up a instead of making a legal AI tool, they make a AI-native law firm and they're selling the work.
1:52:20
I could still go to a chat GPT, a Claude, a Gemini and a future version of the model, I can just say like, "Hey, I need to do this thing. Can you help me do it?"
1:52:29
So, there's still going to be that competition.
1:52:34
But, you're saying that's a much better uh sort of competitive environment to be in than just selling like the software that enables somebody to do the task.
1:52:41
Is that Is that roughly right or do you Is Is there something I'm missing?
1:52:47
Yeah, I'll I'll just add one framework that I think is helpful for people to clarify what all that means.
1:52:53
Um basically everyone is conflating the model's intelligence with what human are really good at, which is judgment.
1:53:00
Uh and I think if you strip those two concepts, then you really understand what the future will look like.
1:53:06
And so, right now the models are becoming really good at everything intelligence related.
1:53:14
So, think of tasks that are um you know, rules based where just logic gets you to the goal, right?
1:53:19
Whereas the judgment is everything that humans really spike at, which is more un harder to define, right? It's instinct. People call it taste. Um call it experience.
1:53:30
And those things are things we think the humans will will remain good at for very long time until intelligence is able to absorb these skills into it.
1:53:42
But it's just really hard, right?
1:53:42
Cuz we we give the models a lot of credit for their intelligence, but they're just trained on set amounts of data.
1:53:47
Um and so, we actually think that this hybrid motion of coupling the intelligence from the model with the judgment from the humans is going to help you build the the next generation of companies, which we call these autopilots, right?
1:54:02
So, a true example, in the law firm, you will automate a lot of the intelligence work that's very much codified, um which is what the model will do.
1:54:13
If you put the employee at the center, someone who's really experienced, who can talk to a jury, who can you know, mark a pause when uh at the at the right time, those are all the things that you learn only from real world experience that a top lawyer will be able to experiment, but that the models might will omit today. Does that make sense? Yeah.
1:54:34
I still I still I like you know, uh may maybe I'm maybe I'm too AGI-pilled, but I but I just assume [laughter] that a sufficiently advanced agent would be able to uh capture and accumulate, you know, the ability to >> [laughter] >> Just talk to Dr.
1:54:50
Alex Wisner-Gross for 30 minutes and he's like, my I'm uploading my brain right now.
1:54:57
No, it's it's it's not that, but it it's just it's just the idea of like a a you know, you put a smart person into an organization where they don't have any experience and they start doing things and they start getting feedback from other humans in the organization and over time they accumulate enough
1:55:12
knowledge and domain expertise that they're able to excel at something that even even using even using the the customer service example, you know, you can you can you've always been able to outsource customer service and you can go work with a firm that is excellent at customer service. The
1:55:29
The alternative has always been to effectively hire a really great leader for that part of the organization, have them build out a team and a structure and an approach and execute uh something that that would be a similar result.
1:55:46
There's like a cost trade-off, but It does feel like overall there's like maybe some pressure on the B2B2B model or like the B2B2C model.
1:55:56
or like the B2B2C model. If instead of uh to use your example of of customer support in uh there was a time when when software companies would sell phone tree software to not not to companies directly to do customer support, but to customer support outsourcing agencies
1:56:17
that would then deploy that and now you're seeing more companies go directly to like like the the the the firm that makes the the the the customer support experience and and go directly to the company that needs customer support as opposed to having this like middle level. And it almost makes me feel like
1:56:34
And it almost makes me feel like we're we're maybe going to rotate even a little bit further into D2C.
1:56:39
When I think about law, there's a lot of stuff where the software would go to the lawyer, the law firm goes to the person and if the person can go straight to the law firm or if the the the legal model, you you wind up with a little bit more B2C activity, a little bit more company directly to to the the service provider or the technology provider.
1:57:03
But but again, this feels a little bit dependent on specific market dynamics.
1:57:08
I'm I'm interested to know your reaction to that, but also what does hold in your mind in the the near-term AGI scenario because there's been a lot of debate over how well marketplaces perform or are network effects still super sticky in a world of AGI or what what are the what are the what are the powers the seven powers >> I guess part of it is kind of close out my thoughts.
1:57:32
Like I I agree I agree with the entire thesis and like the relative short-term like the one one to two years, but then beyond that there's still some uncertainty There always is, yeah.
1:57:44
around like I I don't think it I don't think it solves the the sort of discount that even traditional SaaS companies are getting Sure. Sure. Sure. Yeah, Julian.
1:57:54
On on that point, I think you're you're right then basically what we realize is no one really knows, right?
1:58:02
I've been in many partner meetings recently where everyone just is astonished at how quickly things are moving.
1:58:10
I've been doing that for 10 years and it's just by far the fastest rate of change I've noticed, but some of my partners, you know, Alfred Lin, Pat Grady, they've been doing that for a lot longer and they also acknowledge that everything is moving so quickly.
1:58:24
And so what we're trying to figure out is also not what is going to change, but what are the things that will remain true no matter what, right?
1:58:33
Customers will always want things faster, better, cheaper, right?
1:58:35
And and so that's kind of the thesis that we're describing.
1:58:39
They're just categories that are require more intelligence than judgment.
1:58:44
And so we think that in the immediate the ones that require more intelligence will will be actually really interesting for these auto pilots because you can have you know instead of having 10 humans in one AI, you have one human and 10 AIs, right?
1:58:58
And just that that ratio is just going to shift as the the models get better and are able to absorb what we used to call judgment from the humans that are is turning to to intelligence.
1:59:09
And just to put like a practical framework to that, you mentioned outsourcing.
1:59:13
It's it's actually the best wedge right now is just to focus on outsourcing if you want to capture the the this services spend because there's three reasons.
1:59:23
The first one is that outsourcing it has already been agreed to a third party.
1:59:28
So it's very very easy to sell, right?
1:59:30
Because it's already a vendor that's external to the company. >> Yeah.
1:59:35
The second is that there's an existing budget that's assigned to it which you can peg yourself against.
1:59:39
And and the third is that it's actually very straightforward to swap a supplier.
1:59:45
It's not so straightforward to reorganize your company, right?
1:59:49
We all talk about head count reduction, you know, in terms of ROI, but actually putting that in practice takes time whereas swapping a vendor is very quick.
1:59:56
So I actually think that and that's the prediction is that the the companies that will succeed very well with this autopilot model will start with where there's a lot of labor spend that's being outsourced today.
2:00:09
We mentioned customer support.
2:00:11
We'll mention also things like accounting, tax, insurance, and and so on so forth.
2:00:21
Is there a world where in the next five years or something like venture capitalist truly shift away from software and go back into rare metal mining or something crazy like that?
2:00:35
Like I just keep I just keep going back to like the like Daniel Gross now at Meta, you know, venture capitalist.
2:00:43
I think one of his first AGI trades, the first bet was like in the world of AGI, is copper mispriced? And it was.
2:00:48
It doubled in price over two years.
2:00:50
And I'm just reflecting on this idea that like if the job of venture capitalist is to back software companies, that's one thing.
2:00:58
But if it's if it's just to allocate capital um in in a way that can have uh exponential upside and in a single investment return of fund.
2:01:08
>> would it ever make sense to stop stop investing in software?
2:01:10
I would hate that because we obviously love software.
2:01:15
But uh it it has been funny.
2:01:16
A lot of people have been pushing like, you know, the models are so good.
2:01:19
We just need better harnesses. Yeah.
2:01:21
And to me this is giving me flashbacks to early GPT days where people were like chat GPT is really powerful.
2:01:30
You just need the right harness. All these like wrappers. >> Yeah, yeah.
2:01:34
That had explosive growth very momentarily until and and so I do I Most AGI pilled person ever over here.
2:01:43
>> [laughter] >> I love it.
2:01:44
Uh but but but what what what do you think about about like you know, the the more physical aspects of of the job?
2:01:54
Rare metal mining for example.
2:01:54
Does that does that seem like something people will be working on a couple years?
2:02:00
Yeah, I mean every everyone's talking about the fact that energy will be you know, the biggest bottleneck, right? To train these models.
2:02:05
So it it's very natural that you would go there.
2:02:10
Personally, that's not really where my skill set is.
2:02:11
So I really hope you're wrong and that we'll find other [laughter] areas in between.
2:02:18
I think the real path will be somewhere in the middle. Hopefully, right?
2:02:23
And my my my way of doing things has always been to focus on talent.
2:02:26
And so right now I'm just always asking the smartest people I know, um who do you know who's the smartest person?
2:02:33
What are they working on? And you're you're right.
2:02:35
More and more people are going more into the physical world.
2:02:40
But right now that probably means things like robotics and things like that.
2:02:45
People are making a lot of money trading commodities, but that's that's a different world.
2:02:50
>> [laughter] >> Who knows?
2:02:50
Maybe that's the job of the venture capitalist in 2030 just trading copper all day long.
2:02:57
You [clears throat] mentioned using AI to build a website despite maybe skills that have languished since your time in software or hardware, but How are you using AI tools these days?
2:03:10
Are you still doing mostly like knowledge retrieval, market research for your job in venture capital, or have you actually built tools that have accelerated certain things?
2:03:18
Have you automated any of your workflows?
2:03:20
I'm just kind of curious how people are using AI these days.
2:03:26
Yeah, well, I I I don't know.
2:03:26
I hope my competitors are not listening.
2:03:28
So uh >> [laughter] >> They are.
2:03:32
So you can say I haven't touched it actually. It's useless. Yeah, yeah, exactly. Exactly. Stick to organic Yeah.
2:03:38
organic capital and paper.
2:03:41
>> paper actually is underrated. Exactly.
2:03:43
Back to the basics, pen and paper, and don't try cloud code.
2:03:46
It's it's really really great.
2:03:50
Now jokes aside, it's been really a transformation and particularly the first quarter of this year.
2:03:57
I I mean, I use it for all kinds.
2:03:57
You know, the work is basically sourcing, picking, winning, company building, and harvesting, right?
2:04:04
Then I really have workflows that fit all these boxes.
2:04:09
So it goes from things as, you know, at the end of the week it will just like scan my entire calendar, look at all the meetings I took, look at the granola notes, which is a transcript, do an analysis, and and help me, you know, try and find the key questions in the due diligence.
2:04:23
And you know, it'll help me just prioritize, right?
2:04:26
Because this job is just so much about prioritizing. It's a power law, right?
2:04:32
There's only three meetings in the year that will really that could change your the course of your career and and so making sure that you're a tune in those meetings is really important.
2:04:42
So so that's really what I would say pod has helped me do right now and I'm sure you know open AI will release something next month that will be amazing and so on.
2:04:52
I just have to experiment right then I think I think that's also like what we've seen with things like open claw being so top of mind is like this culture of experimentation is not just in software for engineering now but it's it's transpired across the entire knowledge workforce and you know venture capitalist are not an exception. I love it.
2:05:13
Well, thank you so much for taking the time to come share with us. >> meet you.
2:05:16
Have a great rest of your day. >> Thanks guys. We'll talk to you soon. Goodbye. Take care.
2:05:19
Let me tell you about 11 Labs.
2:05:21
Build intelligent real-time conversational agents.
2:05:24
Reimagine human technology interaction with 11 Labs.
2:05:26
You know, I if I was a venture capitalist, open claw every morning at 5:00 a. m.
2:05:31
text me the name and market cap of the company that I passed on 3 years ago.
2:05:36
I need inspiration to grind harder.
2:05:38
That's the secret to success as is Lambda.
2:05:40
Lambda is the super intelligence cloud building AI supercomputers for training [music] and inference that scale from one GPU to hundreds of thousands and without further ado, we have Owen McCabe the CEO of intercom.
2:05:53
Uh a huge supporter [music] of TVP and as well. Welcome to the show. How are you doing? >> Great to see you.
2:05:58
>> Great to see you again.
2:05:58
Great to see you both also been too long. It's been too long.
2:06:01
>> Man, you have you just have a voice for for for radio live streams, podcasts It's it's it's incredibly calming. I love it. >> love it.
2:06:11
Okay, well what would you like me to say?
2:06:12
I can record a meditation [clears throat] for us. We can Ooh.
2:06:14
Can I start at the hour that one? Do you have time?
2:06:17
Do you have time ASMR funding announcement.
2:06:20
>> time to meditate these days?
2:06:22
It's kind of the only thing that keeps me sane. That's good.
2:06:25
Uh well, you have some new funding for new project. Break us down. What happened? What what's the plan? How much did you raise? Take us through it.
2:06:34
Yes, we raised $250 million. >> Let me hit the gong. Congratulations.
2:06:40
>> [groaning] >> Did that gong get bigger since I was here? It did get bigger. It did get bigger.
2:06:46
It would be cool if it got bigger every every single season, right? Fast take off. Just wait.
2:06:50
There's a fast take off in >> fast take off in gong size over here.
2:06:55
It's growing exponentially.
2:06:55
Except this was actually I think the largest we could find in the US.
2:06:58
So There's always bigger.
2:07:01
I kind of have gong envy right now. I wish I had a gong. We'll send one over.
2:07:04
Let's talk about that later. We'll we'll talk. We'll talk.
2:07:06
So $250 million, what's the plan? Yeah.
2:07:09
Yeah, so I mean, we've been building this thing called Fin for 3 years now.
2:07:15
Service [clears throat] is just this obvious place where AI was going to you know, blow up.
2:07:19
And I was in there and also started to chase it.
2:07:21
We were about a year ahead every everyone else.
2:07:25
And we started doing you know, customer queries and then more complex problems.
2:07:31
Originally, it was answering questions for customers and it's fantastic, you know, the typical response rate with a human was 1 to 2 days.
2:07:36
You'd have to deal with the beautiful imperfections of humans.
2:07:39
Unfortunately, these bots are just far more perfect than humans and they answer in seconds.
2:07:43
And of [clears throat] course, they're cheaper, too.
2:07:45
So, we we started to say better, faster, cheaper.
2:07:47
These things are really better, faster, cheaper.
2:07:48
So, that was a tremendous success.
2:07:50
Fin is the biggest service agent in our category, nearly 100 million ARR or 8,000 paying customers.
2:07:56
You know, some of the biggest logos in the game.
2:07:58
Thank thank you here from Anthropic team. Yeah.
2:08:00
And Snowflake, the cool guys like Polymarket and Vent and everyone else are on Fin. Um so, wonderful.
2:08:09
Um but where does it go next?
2:08:09
You know, clearly it's just so obvious as each of these categories develop that there's just more to explore.
2:08:15
There's none of these categories are in any way fully baked. Yeah.
2:08:19
And that's one of the interesting things about kind of all of these AI spaces that we're still kind of pulling the thread and seeing what's going to show up next.
2:08:26
and the next big play for Finn is not just answering service queries but really getting deep into every stage in the customer life cycle.
2:08:35
And so we call that the customer agent.
2:08:36
And this money will primarily bring the customer agent to market.
2:08:41
Okay, so when I want great customer service for a company that I'm running, I want yeah, you know, search the knowledge base, answer the obvious questions, handle the refunds, track.
2:08:52
That's part of the customer journey.
2:08:56
Are there other touch points that are unlocked by AI along that journey?
2:09:01
Absolutely and completely. Yes.
2:09:01
From day zero on the website when the customer has very basic questions and they get more interesting when the business wants to start to qualify and understand who that customer is, when the customer actually then wants to handheld when they sign up or they want to actually engage and learn and meet sales people, they want to book a demo or they want to talk to a virtual sales person.
2:09:21
I mean, that's just the very top of the funnel of the to get into the application, if it's B2B, there's all the onboarding, all the detailed help required to get them successful.
2:09:31
If it's e-commerce, it's actually finding products and getting it into the cart and bringing them back for more shopping.
2:09:38
So, there is really the literally the full customer life cycle of work to be done by agents and our customer service was just the little sliver.
2:09:47
I want to know what's happening off camera though.
2:09:50
Oh, we we just have our next guest walking in into the studio.
2:09:55
Hopefully they're less interesting than me. Uh yes.
2:09:56
Uh I I I I want to talk about the some of the best customer service agents that I've hired have uh sort of wound up in real people have wound up in sales roles almost.
2:10:11
Like they have a fish on the line, they're reeling them in saying, "You don't want to cancel.
2:10:15
We actually this product does exactly what you want and that the fact that it was delivered one day late, we're happy to refund you, but it seems like the underlying issue is that you actually want this product and so this product, why don't we get you over here?
2:10:29
And that's this like sort of different skill set that comes out in the best customer service agents.
2:10:34
How are you thinking about actually driving incremental sales through customer service agents?
2:10:41
Yeah, I think one of the interesting things is that the lot of the work in these functions isn't entirely separate.
2:10:46
We just needed to draw lines between them when we were running human organizations cuz you can't possibly have one person who's outstanding at service, sales, marketing, success, etc. Yep.
2:10:55
Yeah, and that's what makes the SAS market so complicated.
2:10:59
It makes the SAS market maps too complicated.
2:11:02
>> [laughter] >> Totally.
2:11:05
A pain in the ass, but um unfortunately a pain in the ass is coming with AI because I just don't believe you're going to have these separate sales agents, marketing agents.
2:11:15
You're not going to have these things fighting with each other, having different context and information, different memory in different parts of the customer life cycle, different messengers. Not going to happen.
2:11:25
Um there's going to be vendors that will provide one seamless solution across the customer life cycle.
2:11:31
Um and that's a big thing we're passionate about.
2:11:32
And one of the very fascinating things we've seen building this for quite a number of months now and we're going to launch the first part of it next month.
2:11:38
One of the fascinating things we've seen is that when there's an underlying relationship with the agent and then there's actual connection in the moment and then there's clearing paint.
2:11:51
Um the the context is prime for selling and encouraging the user to do something next.
2:12:00
If you're on a website and you get a super annoying pop-up that says hey John or Jordy, in fact they're not even going to use your name.
2:12:08
They're going to say, do you want to learn more about our whatever?
2:12:11
And you're like get out of here.
2:12:12
It's like not the right time or place.
2:12:14
Whereas if you're actively discussing it can ask you questions back and it's so much more effective.
2:12:19
And so, we've seen just these things the service agent when it starts to do sales, for example, and vice versa, they're far more effective.
2:12:27
And so, [clears throat] the real magic from AI across the board, and of course this agent world, is going to be when the agents can do things that humans can't.
2:12:36
Yeah, one one thing that stands out is like right now you have this spikiness in dialogue as a customer of a business, right?
2:12:43
You might ask a question coming in and it like creates a moment and then it goes away and then you get maybe transition to someone else and you're talking to a real human.
2:12:51
You talk with them a little bit and you get passed to someone else and like there's something about creating, I think, the customer experience of having this continuous conversation with the organization that you're a customer of that I think will be super powerful because it just means that, you know, a split second away the entire time.
2:13:08
Uh I I I was talking with some guys that own a number of like big car dealerships recently and I was talking to them about AI and they weren't fully sold yet cuz they they use like chat GPT, but but they haven't had anything spread across their whole organization.
2:13:27
And I was explaining to them how there's so many different moments where like a customer lands on their website and they want to know if a car is available, but somebody gets back to them too slowly and like they the light was like finally going off for them.
2:13:43
>> [clears throat] >> Do you think that there's a number How many industries out there have been kind of resistant to even using software from Silicon Valley to date that are going to kind of take a second look at all the different tools and things like Finn now
2:13:59
that they maybe have had a cool experience with AI in their in their personal life and they're basically saying they're they're willing to kind of look because maybe the the historical SaaS platforms they tried them, but it wasn't quite uh magical yet. Yeah, there there will be so many
2:14:14
Yeah, there there will be so many spaces, so many verticals that some smart entrepreneurs will benefit from bringing sexy technology to that aren't immediately obvious to those of us in Silicon Valley.
2:14:26
You know, we tend to sell to ourselves. It's really insular.
2:14:30
So, all the names I just mentioned there like Entropic and Snowflake, they're all the cool brands in our world.
2:14:35
However, there's like car dealerships that are you know, I don't know, maybe it's not as big as Entropic, but they're pretty damn big and we can make >> Depending on the car dealership, I I would put car dealership above Entropic and Snowflake personally.
2:14:47
If it's like a Ferrari dealership, Lamborghini dealership, like just on the cool side, we got to give it to you.
2:14:53
>> On on on on the aura factor, sure. For sure.
2:14:56
Maybe not on on the economic factor, but there are there are so many industries that just don't have our attention.
2:15:01
And so, there will be you know, there will be companies like ours selling to many verticals, but really you know, we'll focus on all of the verticals that are adjacent to ours. Mhm.
2:15:12
And there's you know, tens of billions of dollars of opportunity just in that alone. Yeah.
2:15:15
But there will be people who will say we're going to make agents for dentists, car dealerships, or certain government bodies, maybe even for tax bodies.
2:15:22
Like there will be people who will do this because so much of the effort is not just in building the technology, it's actually building the trust as you say Jordy with the actual people who will have to deploy this technology and make it work for them with their legacy systems.
2:15:37
So, there's a lot of very unsexy work that will be richly rewarding for some people who bring to these places.
2:15:45
Talk about the decision to raise debt instead of equity.
2:15:50
Not something you see >> Yeah, I mean Yeah, I know.
2:15:52
I I just realized I had this kind of mhm blind spot for this as an option for for capital.
2:16:00
And there's just so much equity opportunities out there.
2:16:02
We raised a bunch last year for a secondary.
2:16:07
But when you're in our dynamic where you're accelerating as quickly as we are, like every every year we are doubling our our rate of growth pretty much.
2:16:14
There's no way we're going to be able to achieve our the valuation that we're pretty confident getting next year.
2:16:20
And so any equity raise is going to just be massively diluted.
2:16:24
We calculated that raising this in debt is going to cost us about a tenth um the kind of price to shareholders that it would cost in equity.
2:16:36
This this this is, you know, the incumbents have disadvantages.
2:16:37
They tend to move a little slower.
2:16:39
They have older businesses they need to wrangle with.
2:16:43
But they have access to a normal amount of debt.
2:16:46
Like 250 million is a relatively small amount of debt for the size of our business.
2:16:51
Um and I think that you're going to start to see a lot of these older SaaS companies as they pivot AI and they actually find new growth start to lean on debt as an opportunity um to invest rather than you know, super dilutive equity.
2:17:05
And and honestly, I just think it takes a little bit of maturity to take that route.
2:17:09
There's so much pressure from investors >> just maturity.
2:17:13
It's it's it's confidence. >> Yeah. To know your business.
2:17:18
Well, I you know, I like to think so.
2:17:19
But what I the reason I say maturity is that there's so much temptation to go and attain that latest hottest valuation. Your employees want it.
2:17:26
Your previous investors want it. The media wants it.
2:17:30
But if you're able to tell them we don't need it. Yeah.
2:17:35
Actually, it's way cheaper for us to get a lot more capital with far less effort um then you can actually thread the needle in a way that I think's going to work much better for big companies like ours.
2:17:45
>> The gong over here rings just as loud for debt as it does for equity.
2:17:47
So don't let the media >> Yeah, I didn't actually You guys are hiring a ton.
2:17:51
You said uh you're you're bringing on 650 new people this year. >> Yes. Yeah. So that's gross.
2:17:58
I think net it'll be a little less than that.
2:18:01
It's a phenomenal amount of hiring. Um I fear it.
2:18:08
Um I celebrate with great caution.
2:18:11
There are just countless stories of the hubris of CEOs who boasted about how many people they've Yeah.
2:18:15
hired and it's all ended in tears.
2:18:18
So, it's actually not a point of pride, frankly.
2:18:20
Um I think small is beautiful.
2:18:24
But, we are growing incredibly fast and we need all the help we can get.
2:18:27
>> And you've and you've been at this long enough to know the impact of what, you know, a 100 people, 600 people, what what impact that will have on your organization.
2:18:37
So, I I have the full faith in you. Thank you for that.
2:18:42
>> Thank you so much for taking the time to come chat with us. Great to see you. Congratulations. >> to see you both. Take care. Keep doing your thing. Have a good one. Amazing stuff. >> Goodbye.
2:18:50
Let me tell you about Railway.
2:18:50
Railway is the all-in-one intelligent cloud provider.
2:18:54
Use your favorite agent to deploy apps, servers, databases, and more while Railway automatically takes care of scaling, monitoring, and security.
2:19:01
And let me also tell you about Sentry.
2:19:04
Sentry shows developers what's broken and helps them fix it fast.
2:19:06
That's why 150,000 [screaming] organizations use it to keep their apps working.
2:19:10
And without further ado, we have Alex Epstein.
2:19:12
He's the author of Fossil Future live in the TBP NL Freedom. Welcome to the show. Good to see you.
2:19:17
Thank you so much for coming down on short notice.
2:19:19
Um What are your What are your What are your texts been like the last 48 hours? Is it over?
2:19:23
Cuz you're Cuz Cuz John and I were thinking who's our oil guy?
2:19:30
>> [laughter] >> We had one. We had one.
2:19:31
Yeah, and I'm local, too.
2:19:34
So, yeah, if I'm in town, it's great.
2:19:36
I always like coming in person versus Thank you so much for >> better than the other stuff. Appreciate it.
2:19:40
Um So, yeah, a lot's I mean, there's so much in energy, but we can talk about that.
2:19:44
>> lead the conversation however you think it should go.
2:19:46
What are What's going on?
2:19:49
What are What What How should we even be thinking about oil prices right now?
2:19:53
The impact of the war in Iran?
2:19:56
Uh How do we How should we be processing this news? >> [snorts] >> Okay.
2:20:01
So, uh we did this So, there's a million things to comment on.
2:20:05
So, I'm going to try to segment what I comment on, but we did get involved militarily in Iran, which I would just say, in my non-area of expertise, in general, that is a good idea.
2:20:15
I think it's a better idea to have Congress involved if you can get them involved.
2:20:19
And I think it's a good idea to have a lot of expertise on international oil markets being brought to bear >> Sure. when you do this.
2:20:30
And um I think this administration does a lot of good stuff on energy, but I I think it's pretty clear at this point that there was not maximum expertise brought in on this.
2:20:38
And so, the the standard issue when you're talking >> in indicators there would be like the strategic petroleum reserve hadn't been refilled.
2:20:45
>> we we can talk about that.
2:20:45
So, let's let's make sure to get back to that one, but the main thing is just, you know, rough numbers, you know, Strait of Hormuz, which Iran has control over, is 20% of the world's oil production is flowing through that every day.
2:21:02
And so, whether you're looking backward or looking forward, there is no replacement whatsoever for opening that strait and keeping it open. >> Mhm.
2:21:11
And one of the things we've heard from different people is, "Oh, we have a lot of options on the table.
2:21:13
Like, there's a lot of things that we can do."
2:21:16
There are definite things that you can do, but none of them is 20 million, right?
2:21:20
So, it's just I mean, that's that's more than US oil production just flowing through that one place. >> Mhm.
2:21:26
So, there Yeah, that we can talk about the other things, but I think the first thing people need to recognize is if you do not get that thing open, then you have dramatically higher oil prices.
2:21:37
And then, the other thing to get is oil prices really matter. >> Strait of Hormuz.
2:21:40
Dumb question, can't you just go around?
2:21:45
I I mean, you can you can try to some extent, but there's a reason why 20 million a day are being routed through.
2:21:52
>> Because it actually comes out of there as well, right?
2:21:54
>> just I mean all of these things are very, very optimized.
2:21:56
One of the thing about just understanding the energy industry is like things are just very optimized in terms of where the infrastructure is.
2:22:02
You're doing all of these things.
2:22:04
So, it's There are other and we'll talk about this.
2:22:06
I mean there are certain pipelines you can pipe more oil through.
2:22:09
There are alternate routes, but we're talking in the millions a day maybe.
2:22:15
Most of our other options are sort of in the 1 to 2 million barrels a day.
2:22:16
And just so people know barrel is 42 gallons.
2:22:20
So, you're talking about almost 100 million gallons a day for this stuff, but Yeah, so so you don't have a lot of great options.
2:22:25
And I think that's that's very important for I'm going to give some options, but the number one option is keep that thing open.
2:22:33
And there are basically two ways to do that.
2:22:34
One is you just win the war if you want to call it that a lot more quickly.
2:22:40
Like if you get some form of surrender and you have friendly people controlling it, then guess what? It can open. That That's one thing.
2:22:47
So, I'm not an expert on how to do that, but schematically doing that is very effective. >> Yeah.
2:22:53
The other thing which is a little bit >> game game theory for Iran is they actually benefit from global markets being in turmoil because it gives them some leverage over over any type of negotiation.
2:23:09
I mean this is the biggest damage that they can do.
2:23:11
I mean clearly we've seen they're not going to do You're not going to send missiles to Israel and get rid of Israel, let alone do anything to the US, but they have this incredible control of the one of the centers of the world economy and they have a bunch of stuff going on there. So, let's talk about it.
2:23:28
So, option one is you get a surrender and they don't use all of their options.
2:23:31
But then let's talk about their options.
2:23:33
So, they have mines in this thing.
2:23:34
So, they can blow up mines. They have missiles.
2:23:40
Maybe the worst thing they have and this is a change in recent years and I you've had Palmer on here and Ethan on here like the drone thing is just a total game changer cuz you can have a small vehicle, you can launch one of these, you know, Shaw heads out of the back of a truck.
2:23:52
They can go really far and that's a lot harder to deal with than a fixed installation Yeah, they have something like 500 miles of range on a typical Yeah, depending on Yeah, right.
2:24:04
So it's >> drives up to a beach, drone launches as soon as it finds that cargo tanker that has a bunch of oil on it, it hits it.
2:24:10
Yeah, and then and then there's the you're you're worried about that. >> Sure.
2:24:13
That's too long of a way to go.
2:24:13
As as you're going through.
2:24:15
So you have to you have to think about those are the kinds of threats you face.
2:24:18
So how do you deal with that?
2:24:19
I mean basically you need the US to lead some sort of convoy where people have sufficient security and economic assurances that it'll go through.
2:24:27
So what's involved in that?
2:24:29
Well, one thing is if possible you want to get allies involved and it's it's even possible you get unconventional allies like China who has Well, they depend on this.
2:24:37
Now they have a bunch of Unlike us, you want to go back to SPR, they've been filling up their reserves. >> Sure.
2:24:44
They have more reserves relative to their imports than we do relative to ours.
2:24:49
If you think about ours, we have about 400 in the SPR and we can release about 4 million a day.
2:24:56
>> So we don't have as much but but still they really care about this and a protracted thing is bad.
2:25:00
So but for sure Japan, South Korea, India.
2:25:03
So one thing is just at a high level if you're doing this convoy, can we get these countries that are aligned economically, maybe China but definitely these others, can we get them involved because then it's an addition They have military presence there.
2:25:18
That's an additional threat to Iran if they attack one of their ships.
2:25:22
So but that's a macro thing is just can you do this alone or can you do it with allies?
2:25:25
I think ideally you would do it with allies.
2:25:28
Then there's a question of what you do about these these various things and I think the biggest I mean the the other macro thing you can do is just in some way credibly threaten Iran and say hey if you and I can't give the details of how to do it, but if you attack us if you attack in the Strait of Hormuz, it's going to be very, very bad for you.
2:25:46
So just at a high level, can you make that threat so they understand This is going to be very, very No, you know, we've wiped out a couple layers of leadership, so they might be taking these things uh seriously.
2:25:58
After that, I think the biggest thing from what I've heard from military experts is just these drones.
2:26:04
Like, what do you do about these drones?
2:26:06
How do you give any kinds of assurances?
2:26:08
And as far as I can tell, we don't have any one perfect solution, but you can do a certain amount of stuff from the air.
2:26:14
You can do a certain amount of stuff from the ground, although it's very, very expensive.
2:26:17
And then most cost-effective way you can do it is you can take out certain stockpiles and facilities with the proviso these things are decentralized.
2:26:25
And I, of course, have no specific knowledge about where they are.
2:26:28
>> But in theory, if there's the the suicide drone, the Shahed, that's there's a factory and you take out that factory, you have reduced that capacity.
2:26:35
So so that's going to be it.
2:26:35
And you know, you I know you guys had on Ethan Thornton, I'm friends with him.
2:26:38
Uh you know, he talks about this a lot is just this cost asymmetry Yeah. >> issue.
2:26:44
As you just want to be doing things, even if you're way wealthier, you do not want to be spending $3 million to take out $30,000.
2:26:49
Which might be what happening with a Patriot missile battery.
2:26:53
This this this Yeah, that's the kind of thing that happens.
2:26:55
So you need you need a combination of those kinds of things.
2:27:00
And then the other thing, so that's all the security stuff.
2:27:01
So you get the allies involved, you take these actions specifically against drones.
2:27:05
I think the uh what was I going to say about the uh The other thing you can do, I'm just make sure >> just Well, like the other big Oh, the insurance. Okay, insurance. Yeah, the insurance.
2:27:16
So and this this has been floated by the president, and it's it's a reasonable idea, but my understanding is the insurance vehicle we have, the development corporation, does not have enough funding, so you might need to go to Congress.
2:27:26
But basically, you want to be able to say, "Hey, you it's get We're going to lead this as the United States.
2:27:31
We're going to lead it with allies.
2:27:32
If If does anything, it's big trouble for them and we can counter their specific attacks and if something goes wrong, you're insured.
2:27:42
You're you're trying to create that confidence to get enough people going through and then you get a certain amount of stability and then you start to get closer to your 20 million barrels a day.
2:27:50
Either that or you just defeat them really quickly or both.
2:27:53
>> so the other factor is just the loss of production, disruptions to production, plants shutting down, plants being damages damaged. How does that factor in?
2:28:05
Do do those do those do refineries get brought back?
2:28:08
And I'm talking about an allied allied country.
2:28:13
>> You mean just the the trend over time or what?
2:28:15
>> Yeah, generally I mean just seeing seeing the videos coming out.
2:28:17
You've seen places I think like Qatar, Kuwait, places like yeah, basically just saying like yeah, we're going to pause production because we don't even have a place to store this and or they're suffering damages and it doesn't make sense to keep, you know, refineries online.
2:28:33
This is all but this is all just downstream of do you have the route open?
2:28:38
Like if you have the route open, it solves everything.
2:28:40
If you don't, the way to think of the other things is we do have some other interesting options.
2:28:45
We have to think of them as these are temporary stopgaps.
2:28:49
None of like if anyone says, I mean, you hear some crazy things.
2:28:51
People are like, "Oh, Venezuela."
2:28:52
Like we talked about this last time.
2:28:54
Venezuela is less than a million It's just I mean, it's beyond crazy. It's just impossible.
2:28:58
We It's not like Venezuela has the ability to just increase their oil production by any significant amount in the near future.
2:29:04
I mean, we're talking about companies considering going in harm's way to get some incremental boost.
2:29:09
So, Venezuela's essentially useless right now.
2:29:12
>> So so best case scenario in the next year, Venezuela goes from under a million to 2 million? >> No.
2:29:17
I mean, 2 million would be I mean, if we made it our entire goal in life, I don't know.
2:29:22
But but why would you for that?
2:29:24
>> Venezuela was at like 3 million like >> Yeah, yeah, some Yeah, some years ago. Yeah, yeah.
2:29:28
Right, but but there's so many but there's a lot that there.
2:29:31
It's not going to happen.
2:29:33
It's just so Yeah, so Venezuela I would not even That's just in the category of It's not even a bad idea. It's just not an idea.
2:29:37
There's nothing There's nothing There's nothing there. >> No, no. It's right.
2:29:42
So, the bad the bad ideas, by the way, are The worst idea imaginable is ban oil exports from the United States, which I have heard floated in conversation.
2:29:52
>> Seems somewhat reasonable. I want the oil. Okay.
2:29:55
>> oil, so we don't want to sell.
2:29:55
So, I mean, so many different things, but one obvious one, not obvious one, but it's obvious if you know how these things work, is the refiners in the US do not match up well to the oil produced in the United States.
2:30:08
The The refineries in the United States are based on heavier crude.
2:30:11
Like, crude is rated in terms of weight and sourness. >> Okay.
2:30:16
>> And so, our refineries are mostly for heavier crude.
2:30:18
It's actually one reason why Venezuela has some appeal.
2:30:19
Canada has a lot of appeal cuz >> Cuz they have heavier crude.
2:30:24
>> Yeah, and we're going to talk about Canada.
2:30:25
>> good at refining it, but we produce a lighter crude.
2:30:28
>> Yes, because the shale revolution was just this dramatic shift very quickly in oil production.
2:30:33
So, our refineries are not primarily equipped.
2:30:34
That's why when we ended the crude export oil I forget it was 2014 or something like that.
2:30:38
It was just this huge unlock. >> So.
2:30:41
>> Cuz it allows the global market It allows us to produce for a global market.
2:30:45
>> You can't look at American oil production as a monolith.
2:30:48
>> Yeah, it's not It's not fungible.
2:30:50
>> heavy crude and export light crude. Is that roughly correct? >> roughly correct.
2:30:53
Yeah, so that's one thing is it's just insane to do that.
2:30:57
But, in general, think about it.
2:30:57
What is the value of higher prices?
2:30:59
Is it stimulates production.
2:31:01
We're going to talk about all these different ways where we want to unlock oil.
2:31:04
So, you want to tell American oil producers, "Hey, we're going to actually totally screw you over.
2:31:08
It's actually going to be worse for you maybe than it was before cuz you don't even have a market."
2:31:12
So, you're going to strand all this oil.
2:31:14
So, that's that's terrible.
2:31:16
There's also this financial manipulation idea which has been floated.
2:31:18
Everything is terrible if "Hey, let's manipulate the financial markets basically in a way that we're you're selling future short on a certain time period, and then you're buying them and what you're trying to do is lower the near-term price >> Fed policy or Treasury, yeah.
2:31:33
>> you're >> American way to try to solve a world problem.
2:31:37
>> People that like that.
2:31:38
>> The only solutions involve unlocking oil in one way or another.
2:31:41
You do not want to screw up the markets so that people are less inclined to unlock oil.
2:31:45
Nor do you want to screw up the markets to destroy information.
2:31:48
So, again, the main unlock is reopen that straight.
2:31:52
There's no way around that.
2:31:54
But then we can talk about other things.
2:31:56
So, one one category is what you call spare or emergency capacity. >> Okay.
2:32:02
And this this is there's some uncertainty here, but it >> Quickly, this is separate from the strategic reserve? >> Yeah, yeah.
2:32:07
So, well, emergency capacity is the strategic reserve. >> is.
2:32:10
So, spare means you could be you could pretty easily be producing more per day, but you're not because you're not happy with the price.
2:32:17
And this is mainly Saudi Arabia is is considered exhibit A in terms of amount of this.
2:32:23
Now, there's a debate among experts about how much they have. >> Yeah.
2:32:26
But, you know, some people estimate they have 1, 2, 3 million barrels a day of spare capacity they could ramp up on some kind of >> And do they know how much capacity they have and they're just not sharing it or does no one know?
2:32:36
I mean, that's always been the story.
2:32:37
That's been the story of the Gulf, right?
2:32:39
They're not going to If they just flooded the market with every single barrel that they could possibly produce, they would be getting a worse a much worse price.
2:32:47
>> and there's this cartel arrangement that does I mean, in a sense, everyone has this on different time scales.
2:32:52
You think about US shale producers.
2:32:53
If If prices were sustained at $100, guess what?
2:32:58
There's a whole bunch of shale deposits that they could produce at $75 a barrel and make a fortune on that they're not going to do 4 weeks ago when it was around $60 a barrel.
2:33:06
But But in terms of the thing about the Saudi oil is it's produced at a much lower cost.
2:33:10
So, this is them saying, "Hey, we have some of this on the table."
2:33:15
And then the question is you need to be able to produce it, and then you need to be able to transport it.
2:33:20
So, you have estimates around maybe you could get 1 to 2 million barrels a day. So, that's one of them.
2:33:27
In terms of emergency, if you look around the world, you don't have a lot of immediate spare capacity besides that.
2:33:31
So, we have what's called emergency capacity or strategic petroleum reserve.
2:33:35
You know, it's it's Jordy mentioned that we didn't fill up our strategic petroleum reserve.
2:33:38
I think this is unequivocally a mistake.
2:33:40
I mean, if you look at the whole thing of the strategic petroleum reserve is you want it for when you need it and the best possible time to fill is when prices are low.
2:33:51
And is that physically a place with a bunch of barrels that we >> There's a bunch of different parts of the >> Knox of oil.
2:33:57
>> Well, there's multiple different places.
2:33:59
But but in you can think of it like >> We basically stash the physical oil in America somewhere we own it or can access it and it's sitting there going unused until we're Yeah, now there's issues of Biden Biden definitely misused it.
2:34:13
So, they they used it to basically have lower gasoline prices during midterm elections.
2:34:18
Even though parenthetically their entire policy goal was get rid of fossil fuels, which means you want the price to go up so people can't afford it.
2:34:25
It's a very cynical kind of move.
2:34:27
They did it in a way that degraded the facilities.
2:34:29
So, let's say we're at about 4 million barrels a day we can get out of that thing.
2:34:32
We have 400 million in there.
2:34:34
We should have over 700 million.
2:34:36
We could have easily filled it up.
2:34:38
I mean, it was a perfect time when you're talking about $50 barrel oil, $60 barrel oil.
2:34:41
Now, it take And and can I do the basic arithmetic of 4 million a day, 400 million in the reserve, 100 days of oil?
2:34:52
But that's not actually if we're talking about like relieving price pressure, you could you could trickle it out over a year and have 120 or 1/4 of the effect.
2:35:02
>> But just think about 100 million dollars a day 100 million barrels a day is roughly the global market. >> Okay, okay.
2:35:06
>> So, so 4 million is just that's its maximum capacity. >> Got it.
2:35:10
It's kind of analogy to batteries.
2:35:11
We're talking about batteries.
2:35:12
So, there's a certain amount like with a battery if you hear there's a 4 gig there's a 1 gigawatt battery installation.
2:35:17
That means that's the It usually means there's 1 gigawatt for 4 hours.
2:35:22
But, if you wanted to do half a gigawatt, then you could do 8 hours, right?
2:35:26
It's the same deal with this.
2:35:28
But, keep in mind, this is not 100 days of US oil demand.
2:35:30
And by the way, for the refinery reasons, we can't just supply it all with our SPR.
2:35:34
It's just the maximum output is But, 4 million barrels a day is something we're That's 1/5 of what's flowing through the Strait of Hormuz.
2:35:43
So, if Yeah, it's It's a thing.
2:35:46
But, but yeah, you can think of yeah, maybe we'd be willing to do 1 to 2 million barrels a day for hours.
2:35:49
Now, interestingly, the International Energy Agency, we are part of the International Energy Agency Reserve Program.
2:35:55
So, we have about 400, but they have about 1.
2:35:58
4 400 million, they have 1. 4 billion.
2:36:01
So, we have allies around the world that can also release this oil.
2:36:06
So, maybe we could get 1 to 2 million from them.
2:36:08
Now, last At least last I checked the news, they hadn't agreed to do this cuz they don't think it's an emergency. Sure.
2:36:14
But, that's when it's there for. Yeah.
2:36:16
So, notice the pattern is what What point would they think it's an emergency? I don't know.
2:36:20
I mean, and we, you know, we can put pressure on them.
2:36:22
I I think I think some of these determinations have to be made in conjunction with the military determinations.
2:36:28
You can't make them in isolation cuz if you think about what's the military objective, how close are we to that?
2:36:34
How or how close are we at least to opening this up via high security and financially backed convoys?
2:36:41
Then you can think about the timetable here.
2:36:44
If you think about well, the Strait of Hormuz is going to be closed for 5 years, let's just get used to being poor. Black bill.
2:36:53
No, I mean, this is People don't get oil.
2:36:54
Like, oil is There's a reason why I focus a lot of my life on oil.
2:36:59
Like, there is nothing more There's no material in the world of energy that is more valuable than oil because oil is just so unique in terms of it has this very high energy density and portability.
2:37:09
And there's nothing like it.
2:37:12
nuclear has higher >> plutonium. Yeah, yeah.
2:37:13
Doesn't have the same portability. >> going to be rich.
2:37:16
We're going to have energy too cheap to meter in 5 years.
2:37:19
Yeah, I talked to a lot of nuclear founders.
2:37:20
Well, I I Look, I love nuclear as much as anybody.
2:37:23
They don't They don't even have a portability solution though for the near future unless you're talking about aircraft carrier or an icebreaker or something like that.
2:37:29
>> So, I'm just saying the world runs on mobility and oil is ideal for mobility.
2:37:34
So, Oh, mobility in the sense of like using nuclear to actually power trade and commerce and moving ships around. Yeah, yeah.
2:37:40
I mean, we Even if you can power the grid, you will be poor because you will not be able to trade.
2:37:43
Yeah, trade is really im- important. >> important.
2:37:48
>> Trade is Trade is And well, and also it's within the country, too.
2:37:49
We don't have the nuclear trucks to move things around. >> Sure. Sure.
2:37:53
Yes, I see you obviously want super cheap electricity and you want super cheap transport fuel.
2:37:56
And for And And we want to see if we can electrify some of this stuff. >> Yeah.
2:38:02
If you can do it cost-effectively, it's just >> working on like electric bulldozers, for example.
2:38:07
It's very hard and And then yet by the way, Yes, what you need a lot of oil to even get that whole supply chain started cuz you need the mine.
2:38:14
>> Anyway, bottom line, yes.
2:38:14
I I meant that thing about we can all be poor very seriously. >> Very seriously.
2:38:19
So, we have If you look at the spare capacity, I think that's all of them.
2:38:23
So, we have the I the rest of the IEA, we have ours. Mhm.
2:38:24
And then to my knowledge, we have Saudi {slash} UAE.
2:38:28
Them using spare capacity and then piping it Mhm. where they can.
2:38:31
So, maybe combined we're talking about 6 million a day. Mhm.
2:38:36
If you use all of those, I mean, somebody could imagine 10 million a day.
2:38:40
And all of these have time particularly the emergency reserve things, these all Like the at least the Saudi one is you're sort of you have a deposit a big deposit.
2:38:50
Here, it's just you have a small reserve.
2:38:51
So, you can't do this forever.
2:38:52
And of course, there's a risk of if you deplete the SPR more, what if you really really need need it >> Yep. in the future.
2:38:59
And then there's a bunch of nearer term stuff that you can do.
2:39:02
And some of which I like just cuz it could be done anyway.
2:39:05
So, number one to piss a whole bunch of people off is let's stop using this Jones Act.
2:39:08
Do you know what this Jones Act thing is?
2:39:10
I've heard of it, but explain.
2:39:12
So, the Jones Act is a set of restrictions that say that you can only transport things among ports in the United States if you have everything is basically American.
2:39:22
So, it's owned by an American, it's run I don't know all the It has all these >> basically American ship, American crew. >> Yes.
2:39:30
So, it's it's It leads to all these crazy crazy things like we end up importing things from like instead of importing it from one part of the country to another you end up importing it from some really far away place.
2:39:39
It It leads to that kind of thing.
2:39:42
But, in this case in particular, this this as Californians we should all recognize this.
2:39:48
We don't have oil being piped into California. Yeah.
2:39:50
So, we want very efficient maritime transportation and if you suspend the Jones Act then you can get more efficient transportation which means we can at least have lower prices here.
2:40:01
And we can have less of a price shock. >> Mhm.
2:40:04
But, in general I want people to get used to having no Jones Act because I want there to be no Jones Act.
2:40:07
So, that's That's a good thing to do. Mhm.
2:40:11
Uh my favorite idea here is >> And And the the I imagine there's a bunch of powerful lobbies that will fight and kill for the Jones Act because there's an entire industry built around uh Yeah, they have this idea that we can only have a good shipbuilding industry.
2:40:28
I don't know what people think is happening with our shipping like our ship building industry is not going as well as other people's.
2:40:33
>> It didn't exactly work.
2:40:35
Yeah, I mean it worked to protect certain people's >> I'm just saying if you look at American shipbuilding, it's it's certainly hasn't allowed us to be competitive on on a global stage.
2:40:45
>> The shipbuilders would say we're early, but I take your point.
2:40:49
Early, but you've seen But, I'm just saying [laughter] like how many decades of declines Yeah, okay. Well, do you need?
2:40:54
I I think John's like we're still early.
2:40:58
>> It's like a smiling curve like churn.
2:40:58
It goes down and then goes back up.
2:41:01
So, let's let's That's that's the bull case.
2:41:04
That's that's the All right, well, let's Yes.
2:41:05
It's definitely not going to help us right here.
2:41:09
So, so the Canada here's the most interesting thing, which I is a pet issue of mine because we're we're like this administration, I should say, does a lot of good stuff on energy.
2:41:16
I say that I, you know, I'm in a fortunate position where I get a lot of people ask for advice and stuff like that.
2:41:22
I feel like one of my jobs is to tell people things they're not doing.
2:41:25
A lot of things they'll just do right on their own.
2:41:28
But But Canada is not one of those things right now.
2:41:30
We are just absolutely sleeping on Canada as an opportunity.
2:41:34
So, Canada has no people and infinite resources and is really friendly. Yeah.
2:41:40
So, you do they have these oil sands, you know what these things are? >> explain.
2:41:42
So, so they're just like these deposits of oil.
2:41:44
It's like nature basically committed an oil spill. Yeah.
2:41:48
Is the way to think of it.
2:41:48
So, there's all these oil sands.
2:41:50
For years, they were not very useful.
2:41:52
>> viable because it's not just you drill down, black gold comes up.
2:41:54
It's like buried in the sands, needs to be refined.
2:41:58
>> There are different kinds of things.
2:41:58
You can mine it out directly or you can do what's called in situ, which is you heat it up underground. >> Oh, and it liquefies.
2:42:05
In any case, we they have just this unbelievable oil seep.
2:42:10
They have way more oil deposits than we do, way more oil reserves, but they have on the order of a third of the production.
2:42:16
And in part cuz they have absolutely terrible policies, which are their fault, but in part we have had terrible policies, including getting rid of the Keystone XL pipeline and this kind of thing.
2:42:25
So, but if you just think of Canada, it's just we should be so they have water.
2:42:29
I mean, again, no people, 1/10 the population, infinite raw materials. They got everything.
2:42:33
They got timber, they got natural gas, they got oil.
2:42:38
Like And we're talking about Venezuela, they're friendly country, right?
2:42:42
They got smart people there.
2:42:44
They're They haven't been driven out of the country cuz they're afraid of getting killed by Chavez or Maduro or whatever.
2:42:49
So, this is an opportunity to say, "Hey, Canada, let's work let's have a task force.
2:42:54
You guys are at a low in terms of we don't have pipelines, but we can do rail.
2:42:58
At least we can do rail, maybe trucks, but at least rail.
2:43:00
Rail transport from Canada are really low right now.
2:43:04
So, let's have an initiative where maybe you bring a couple hundred thousand depending on your expectations, maybe we can add a couple hundred thousand barrels a day.
2:43:11
So, that's one that excites me cuz I want to be like the US Canada superpower thing is this huge opportunity that we're sleeping on because of the idea of oh, we should only be doing it in {quote} America. Oh, interesting.
2:43:26
>> By the way, unbelievable.
2:43:26
I think they have the best uranium in the world. >> Yeah.
2:43:29
It's uh Maple syrup's up there, too.
2:43:32
>> Yeah, I don't have a Canadian personality, but I love that place.
2:43:35
>> They're strategically It makes a lot of sense.
2:43:36
Strategically and they're very friendly. >> Yes. They're very friendly.
2:43:39
>> Maybe we should be friendlier. Yes. You know what?
2:43:41
You know, we should and we have trade agreements coming up, so there's that.
2:43:43
The other thing is we have some, you know, we can increase capacity here.
2:43:48
As I mentioned, depending on prices going up.
2:43:49
So, there's that, but that might be a couple hundred thousand barrels a day. Mhm.
2:43:54
But so, high-level, open the Strait of Hormuz, that's the only thing that really works.
2:43:57
Then you can talk about Saudi Middle East spare capacity, our emergency capacity, and IEA emergency capacity, and then below that you have Canada, Jones Act, and US >> That's great. >> uh production.
2:44:11
>> That's a really, really thorough list. Thank you. I appreciate that. Anything else, Turi?
2:44:15
What are you uh what are you going to be paying attention to most closely uh while we've been >> chart?
2:44:20
Is that important to you? Well, it's important.
2:44:22
I mean, I I'm It's weird.
2:44:25
I work with uh I run like a kind of a network of energy executives, but I'm focused on the policy side of things.
2:44:32
So, I'm often laughably out of touch.
2:44:34
I mean, I know in general within $10 usually what it is. Yeah.
2:44:38
But I as a policy, I don't I don't invest in energy.
2:44:42
>> at longer term, 5 years >> But but also like I'm deliberately avoiding exposure to the commercial part of it because >> Yeah, you want to be independent.
2:44:49
>> Yeah, I would I want to be independent, so yeah, I'm not but at the moment, yes, I am looking I am looking at the prices.
2:44:55
But but I'm I don't do predictions, but I'm just saying I think there's a service right now in just laying out the options for people cuz even in the government, I mean, you know, not everyone has a lot of expertise in in oil and this is the kind of thing where yeah, it's kind of why I think you texted me this morning.
2:45:11
I mean, sometimes you need to know sometimes you need to know about oil today and sometimes you don't.
2:45:15
But when you need to know, you need to You know me. You really need to know. >> Yeah.
2:45:19
Well, it seems like whether you're in the business community or just, you know, humanitarian, uh everyone should be rooting for a swift and peaceful resolution.
2:45:29
>> Yeah, do you do you have any insight?
2:45:30
This this might be totally out of your wheelhouse, but insight into how uh how people running businesses that are most impacted by fluctuations in oil prices?
2:45:40
Like are they were they have they been actively hedging over the last couple weeks?
2:45:45
Are they doing like how how do you run your business when you're in input costs can go up dramatically, potentially at a much higher rate than you can adjust your own prices.
2:45:53
I mean, you just in general, these guys are sophisticated.
2:45:59
I think within the bounds of doing it.
2:46:02
I don't think you can hedge indefinitely for this big a price Yeah. kind of differential.
2:46:07
Like somebody is going to say I mean, but it's it's also you think about there's on the consumer side, but on the production side is is even more dramatic if you just look at the margins of these guys.
2:46:17
I mean, the oil industry is just just fascinating.
2:46:20
I mean, I just remember anecdotally when I I came up with my first major book, The Moral Case for Fossil Fuels, in 2014, even I could see speaking engagements correlated to these prices cuz you had a big price crash >> Sure. around them.
2:46:32
And then you think, "Wow, like Yeah, we don't need we don't need Alex Alex >> variable [laughter] variable Yeah, it might be nice to keep the troops motivated, but you know, I'd probably rather be employed than but it's it's a hard it's a hard business and one of these things is you just the people who survive have a very strong constitution and and they they they can manage the risk very very well.
2:46:58
I was talking with a bunch of these these guys and actually these people we know in common and like some of the better ones some of the two of the young superstars in in the industry they're just saying like like we love the chaos cuz we can just No, we can handle it.
2:47:11
Like when the prices go down. >> Sure.
2:47:14
Cuz that's you think about what they've been dealing with over the last year they've increased production prices have gone down and and you look at this this administration is very focused I would say too much focused on having prices go low.
2:47:25
This is this is one of the variables here is although the administration I don't think was fully was not fully prepared for this situation and in general they are maniacal about prices going low.
2:47:34
Like Trump in his mind has a very strong focus on $50 a barrel which I think he feels like that's always the perfect price for oil. >> Sure.
2:47:45
And the people in the oil industry say wait a second like we had inflation of everything else.
2:47:47
Why does our one product have to stay at $50?
2:47:49
And that's part of the appeal of Venezuela rightly or wrongly cuz it's not going to make a big difference soon but it's like how do we keep oil prices >> low. low.
2:47:58
But that's that's hard to be in that industry as a consumer Mhm.
2:48:01
as as a as a producer rather and it's hard as a consumer.
2:48:05
So they have financial instruments but the main thing is you just you need to you need to just have uh the flexibility in your business in your business model and it's it's oil is again it's the most valuable material in the world of energy which is why everyone uses it but it's also why it's so inelastic and you have these price fluctuations.
2:48:26
So you just need to come to terms with that or somebody needs to figure out something better.
2:48:30
But it's it's really hard.
2:48:33
And of course policy-wise we can have better policy to make it more stable but it's still that thing where it's so valuable that guess what you get a little extra demand a little less supply price goes up a a lot it's it's it's different than other things which you can substitute for much more rapidly.
2:48:50
You're out of a job when we build the Dyson sphere.
2:48:53
Oh man, I got I'm I'm working on this AI thought leader thing we talked about last time.
2:48:56
So I got I got plenty to do.
2:49:00
I got plenty to do with my time. Thank [laughter] you.
2:49:02
As long as we need energy you'll you'll be doing just fine.
2:49:05
All right guys, good to see you.
2:49:06
>> so much for coming on the show.
2:49:06
Let me tell you about Shopify.
2:49:08
Shopify is the commerce platform that grows your business and lets you sell in seconds online in store on mobile on social on marketplaces and now with AI agents and let me also tell you about AppLovin.
2:49:18
Profitable advertising made easy with axon. ai.
2:49:20
Get access to over 1 billion daily active users and grow your business today.
2:49:24
And without further ado, we have our next in-person guest live in the ultradome.
2:49:29
Michelle Vols, how are you doing? Good. >> see you. Good to see you.
2:49:31
Thank you so much for coming down to the TVP in ultradome.
2:49:34
Uh for those who aren't familiar with your venture capital career and your marathon running career, introduce yourself.
2:49:41
Yeah, so uh I am Michelle.
2:49:43
I was previously at Andreessen Horowitz.
2:49:46
I just launched my own fund Pax, which is a $50 million early stage venture fund focused on foundational categories.
2:49:53
So I was on the American Dynamism team at Andreessen Horowitz similar themes to that just earlier stage.
2:49:59
Biggest lessons from Can we hit the bong first?
2:50:03
>> we got to hit the bong.
2:50:03
I think I think Michelle should hit it. >> her to hit the bong? Here you go. Hit the bong. We're here in person. Go for it.
2:50:08
Now it's already sort of broke. You won't break it. >> Smash that.
2:50:20
>> [laughter] >> Anyway, let's let's start with like lessons from Andreessen Horowitz. Like what did you learn?
2:50:27
What are your biggest takeaways?
2:50:28
What are what's the playbook that you think you will be continuing and then we can go into what you think might be different?
2:50:33
Yeah, I mean I think in the categories that I look at it is common for founders to need to raise a lot of capital. >> Okay.
2:50:40
And the ability to be able to fundraise for that capital ahead of maybe what are traditional milestones in software investing is very important. >> Totally.
2:50:51
You need to be able to be a magnet for capital, for talent, and sell a vision because what you're doing is like very hard with with fewer obvious proof points maybe at the early days.
2:51:02
And I think identifying that in founders early is something that Andreessen Horowitz is very good at.
2:51:06
Yeah, in in American Dynamism companies, hard tech companies, it does feel like there's like high CAPEX requirements, high human capital requirements.
2:51:16
You're not going to automate everyone on your supply chain on day one.
2:51:22
Uh and so there is this like how you talk about growth, how you talk about about scaling and success and finding product market fit.
2:51:32
It's it almost feels back to like the okay, we we're scaling DAU and we're let's focus on that metric instead of like profit in software.
2:51:39
What does it look like in hard tech?
2:51:42
Is that the okay, we have a program of record or we have these SBRs coming in or we have this long contract that is very solid, but we're not going to draw down on it for a few years, so we have a capital need.
2:51:55
But how how do you think founders in hard tech should be talking about what progress looks like at at top line or whatever top line means?
2:52:03
Yeah, I mean I think that's the the hardest part.
2:52:06
Like in software, there's such a structured set of metrics that everybody has been trained around.
2:52:10
Like growth rate, like weekly active active users, revenue.
2:52:15
It is so different company by company in these sectors, and that's why the storytelling is important because if you're talking to later stage investors, they might be trying to make a pattern recognition off of a pattern that doesn't quite exist yet.
2:52:27
And so in defense companies like yes, there are government contracts you should be trying to get, you should be working towards programs of record, but there are like different intermediate proof points bef- before that.
2:52:37
I usually explain it to founders as like the three things you're trying to make progress on are uh people, product, and then like revenue or traction.
2:52:46
And so it's like are you hiring the best people and the people that are experts in the different categories that you need?
2:52:52
So you maybe you need a mechanical engineering hire, maybe you need a government sales hire, uh maybe you need software as well.
2:52:56
Um are you making progress on people?
2:52:59
Are you making progress on product that's tangible?
2:53:02
Like can you show that there's like developments?
2:53:04
Maybe that's like you've built something and it's like test like gotten to a test milestone.
2:53:09
Maybe it's in the hands of users.
2:53:09
Like are you making product progress?
2:53:11
And then are you getting some sort of signs from customers that like they are willing to pay for this?
2:53:17
Like are you getting contracts? Are you getting LOIs?
2:53:21
Like are you getting something that shows that there's a commercial use case for this?
2:53:24
I think sometimes in like hard tech there's a million names, hard tech, deep tech, frontier tech, Americanism.
2:53:30
Um I think sometimes people skew more towards thinking it's just deep tech, which is like uh I would describe it as like >> Exactly, like science risk.
2:53:37
Like nuclear fusion is probably like the most extreme example.
2:53:42
Um that's hard like it's like of course there's a customer if you can do this, but like we'll see what the timeline is. Hopefully soon.
2:53:49
Um but I think in a lot of the categories I look at there are commercial proof points you can make.
2:53:53
Like there's usually a latent demand or like or an immediate demand.
2:53:55
It's like how do you show that you can get those customers and they can like uh signal that they will buy from you. Yeah.
2:54:04
There was this viral Satrani article that said software is dead, sell everything.
2:54:10
Uh how are LPs thinking about hard tech as a category with that backdrop of like the SaaS-pocalypse?
2:54:18
I imagine it makes things easier for you, but has it actually rippled through the LP community to the degree that your investors are more excited than ever to have a hedge against chaos in automatable software, all of that. Yes.
2:54:34
Uh I I will say this with love to all of my LPs.
2:54:37
Uh I think founders are typically like leading indicators in categories.
2:54:40
Like they they're at the cutting edge.
2:54:42
Like they're seeing around corners.
2:54:44
Often like great founders will see something before VCs do.
2:54:48
Like VCs are lagging indicators.
2:54:50
Like once something is hypie, it's like usually too late to find a good early-stage investment.
2:54:55
LPs are lagging lagging indicators that uh once something is very mainstream and there starts to be some signs of returns, then they're like, "We should pay attention to this category."
2:55:04
And so I would say uh thing that was more surprising um partly because I live in my own echo chamber and constantly hear about like all of these great companies building in like defense tech, hard tech, and like industrial tech.
2:55:18
Um LPs weren't as familiar as I expected.
2:55:20
So like there was a level of education.
2:55:22
I think there's still some wariness like there haven't been a ton of exits yet.
2:55:26
Um I mean, I think that might change if SpaceX IPOs.
2:55:30
Um it'll be one of the biggest IPOs in ever in history.
2:55:35
Um but I would say like the the appetite for it grew over time as I was fundraising.
2:55:40
On SpaceX, uh yeah, I mean, great point. It's it could be like 1.
2:55:46
75 trillion is the number that's being floated.
2:55:48
It's it's staggering amount.
2:55:50
It's going to return like not just like one venture capital fund.
2:55:54
It's going to return like 20 venture capital funds.
2:55:56
Like everyone who invested at early stage is going to see their fund return from that.
2:56:00
But all the LPs had to go on a 20-year journey with like a million continuation vehicles.
2:56:05
How are you thinking about timeline messaging to LPs?
2:56:08
Because 10 to 12 years is what I've heard recently is kind of standard, but I've talked to some VCs that say, "Yeah, like most of my LPs want their money back in like 6 to 7 years."
2:56:18
And yet in hard tech, it feels like it might be longer.
2:56:22
Are you communicating a particular timeline or is that even a relevant conversation?
2:56:26
>> Yeah, I think like every every fund roughly pitches 10 to 12 years.
2:56:29
That's sort of the expectation.
2:56:33
Um I I think it's actually an interesting Catch-22 because if you had um ownership in SpaceX, your LPs actually like would like hold on to that. >> Yeah, totally. >> Like keep that.
2:56:42
I think it's going to keep going up.
2:56:43
But maybe you would have an option to take some off the table and return some, you know, money to LPs if that's what they wanted.
2:56:50
But often if it's like a really, really good deal, LPs are like, "Oh, I think it's going to keep going up."
2:56:55
Um but I I think it's like case-by-case.
2:56:58
Like I actually just was talking to LPs about this.
2:56:59
Um and they were like, "Sometimes it makes sense and sometimes like hold on."
2:57:03
And I think if you build a lot of trust with your LPs, like they're usually along the ride with you.
2:57:08
What what GEOs are you spending the most time right now?
2:57:12
Obviously the uh it feels like we have so many so many founders that are just like in in our kind of neck of the woods here.
2:57:18
But are you spending a bunch of time in Texas, Bay Area?
2:57:22
Where where do you expect to deploy most of your dollars?
2:57:24
Yeah, I think like there's five GEOs and it's San Francisco, LA, uh Austin, New York, and DC.
2:57:31
I'm actually seeing like a significant amount of companies in DC.
2:57:34
Austin, I would say though is the biggest surprise for me.
2:57:38
I know people have been trying to make Austin happen for a while and I I was skeptical and now I think like it's sort of been an explosion of great companies.
2:57:46
It was kind of the same thing with LA.
2:57:48
Like there was this long LA movement a while ago.
2:57:50
Snapchat was sort of like the first big, like, you know, public tech company or one of them to come out of LA.
2:57:57
And then the El Segundo thing just happened like completely organically without any of those people that were set cheering for LA.
2:58:02
It just sort of popped up and it seems like uh Austin's going through a similar thing.
2:58:07
Uh what's unique about the companies that are building in DC?
2:58:08
Are they thinking more demand generation first, more more work on Capitol Hill first, and then maybe they'll set up a satellite manufacturing office or work with like a contract manufacturer to deliver their capability.
2:58:21
Yeah, I think like you're seeing founders who are working at companies like like Palantir for example, that have a big office in DC.
2:58:29
They've been relocated to DC.
2:58:31
They're close to their customers, and then they spin out, start a company.
2:58:36
Um usually these companies will end up opening an office in New York as well.
2:58:40
Um it's just a deeper pool for engineers, but there's a surprising amount of really strong engineering talent in DC that is like pretty fixed.
2:58:48
Like they want to stay in DC. They have a life. They have a house. They have a family.
2:58:51
Um so it it can be a very good place to start a company.
2:58:54
What's the view on shipbuilding?
2:58:55
Jordy says it's impossible to build a ship in this country.
2:58:59
He was saying it can't be done. Is it possible?
2:59:02
>> We We need to figure it out, otherwise we're going to be behind.
2:59:04
I think I mean like the optimistic take would be like we can make it possible.
2:59:09
Um we should make it possible.
2:59:11
The pessimistic take is if we don't make any changes into the regulatory environment or the cost of building these things, um it might not be, and that would be a shame. Yeah, that makes sense.
2:59:22
Uh strategy with the fund Yeah.
2:59:24
Are these like $2 million lead, pre-seed type checks? What what's the deal?
2:59:30
Yeah, um on average one or two million dollar checks, like early stage, try to be in like the pre-seed, seed, and occasionally a little bit later um rounds.
2:59:38
Um like to partner really closely with founders.
2:59:42
Um and I think the special thing about the early stage, which is just my favorite stage of company, is you get to be very aligned.
2:59:48
It's like a significant check for me if I put, you know, one or two million dollars into your company, and I'm very motivated to help you be successful to get to the next round.
2:59:56
And then for that next round, I can be much more strategic in in helping you get there because I can't lead your next round.
3:00:05
Whereas I think sometimes at the multi-stage funds, it can be a little a tricky to give advice for future fundraisers because you might be able to participate and so you have to be like more delicate about it at a big fund and as a small fund you just get to like be totally aligned with the founder.
3:00:21
One last question about sort of the category targeting uh American dynamism made a lot of sense as like a theme around defense tech and hard tech and you put a manufacturing company in there.
3:00:34
One of the interesting subcategories of AD that I always that always popped out to me was something around education uh like re-educating workforce education also just like education technology broadly.
3:00:46
Do you think that that's within your purview or is that something that uh you're sort of carving out and focusing more on the deep tech hard tech physical products?
3:00:57
Yeah, I would say I like to look at things that sit at the infrastructure layer of society which could include education and housing sometimes even healthcare often things at the intersection of hardware and software or tech and government or just like doing something fundamentally important for like the broad majority of Americans.
3:01:15
Um I think American dynamism is this beautiful phrase that can be an umbrella phrase as well.
3:01:20
Um I I sort of embrace that as well with packs but at the early stage like what you're really betting on is people.
3:01:24
Like there is so little to understand except like the team and a dream.
3:01:29
Like are you building something in a category that's very exciting are you taking the big swing?
3:01:34
Do you sort of have a right to win in that category and some sort of unfair advantage?
3:01:38
I like founders that have like worked in the industries that they are building in before um and do I believe you can be that magnet for capital and talent and everything?
3:01:48
Yeah, where where are the big talent pools these days?
3:01:49
There's the SpaceX mafia the Palantir mafia feels like there's an Anduril mafia now.
3:01:53
Where else is it going to the Basecamp power Basecamp power maybe in Austin eventually.
3:01:59
Yeah, I think Base I think Saronic is starting to see a few things.
3:02:04
Basically any company that's that's like reaching escape velocity or like um like escape velocity and product market fit has grown really quickly and has hired exceptional people.
3:02:14
There are some people that really love that zero to one stage and the company gets big.
3:02:19
Like I think we're seeing it at Anduril. We saw it at Palantir.
3:02:21
We continue to see it at Palantir.
3:02:24
People people learn how to grow quickly and then want to apply that to their own company. Amazing.
3:02:30
Well, thank you so much for coming by. >> Congratulations.
3:02:31
I'm sure we'll have we'll have a lot of your companies on.
3:02:35
>> Let me tell you about Vanta.
3:02:35
Automate compliance and security.
3:02:37
Vanta is the leading AI trust management platform and let's go over to some hard tech that's happening with Stargate.
3:02:47
By the way, John, apparently Trump said something to the effect of I think the Iran war is very complete, pretty much. >> Oh, okay.
3:02:58
Well, so we may be on the verge of new era peace.
3:03:04
>> down to like pre like Friday, pre-weekend prices. >> 80-90? >> 83. >> Okay. 83. Okay.
3:03:10
Well, maybe he was watching.
3:03:13
Maybe he was watching Alex Epstein and Alex said, "Hey, we got to we we we got to wrap this up. We got to stop."
3:03:20
The price is too too darn high, says Alex Epstein.
3:03:23
Oracle and OpenAI allegedly dropped their Stargate expansion lease.
3:03:30
Meta swooped in the same day.
3:03:32
Ben Pula Dean says that's not weakness, that's a hot real estate market.
3:03:36
And then he goes on to comment that the real story is OpenAI passed on six additional buildings because power won't be ready for a year.
3:03:43
The power is is is the gating factor here, not the chips, interestingly.
3:03:48
We are in the power the power bottleneck, the energy bottleneck.
3:03:52
And by then Reuben replaces Blackwell, so why are you going to build a Blackwell data center when you could build a Reuben data center.
3:03:58
So why expand with last gen chips?
3:04:01
The eight building 400,000 GPU base build continues.
3:04:07
This is smart Alex capital allocation, he says.
3:04:10
And so, this is based on reporting from the information about Oracle and OpenAI sort of adjusting the plan for for Stargate.
3:04:17
I am particularly interested in digging into the the Stargate of China.
3:04:27
They have a project that's the Chinese Stargate, funded to the tune of 36 billion, not not up at half a trillion, but still >> very big number.
3:04:38
>> number and it's growing.
3:04:38
And I've been I've been digging into it a little bit trying to learn as much as I can.
3:04:42
We'll do a deep dive probably with someone at SemiAnalysis who knows a lot more about it.
3:04:48
>> chat GPT is back at the top of the App Store leaderboard.
3:04:53
>> we do have to we >> Adam GPT says GPT 5. 4 >> sends its regards.
3:04:57
You know what that meme is from? Big Chungus. It's a good one.
3:05:01
Uh anyway, uh Uh yeah, so so I'm I'm kind of surprised by this for the the one reason being Yeah.
3:05:09
I expected Anthropic's been spending a ton of money on marketing.
3:05:11
They started ramping it up. Yeah, yeah. Yeah, yeah. I see that all the time.
3:05:16
I expected them to basically be willing to spend whatever it took to stay on top.
3:05:22
>> Yeah, but I mean, it's capital war.
3:05:22
So, both sides are going to spend a lot of money, right?
3:05:25
And you know, there was what? What's up?
3:05:28
Wasn't the rise in the Claude like ranking just because of the the DoW news and now that's like not on the front page?
3:05:34
>> Well, so so yeah, I that's where I was getting.
3:05:36
It was like when the Super Bowl ad dropped and we had some harsh words for the the nature of the characterization around ads, our favorite thing in Claude, led us to launching Claude with ads.
3:05:48
But, our question was will this Super Bowl ad move the needle for the consumer app?
3:05:53
Like a Super Bowl ad is a massive campaign.
3:05:56
Millions, tens of millions of people will see this. It went viral.
3:06:00
Like, it's will it work on people?
3:06:00
Will people see this ad that's like sort of deep in the weeds around how ads work in LLMs?
3:06:07
Like, will this be enough for people to go and hit the download button? There's a hurdle there. Will they do it?
3:06:12
And they did it a little bit.
3:06:13
It went I think Cloud went from like like 26th or it would started climbing then.
3:06:18
And then, you're right, the the Department of War narrative really really shot up.
3:06:24
But, uh as with any uh any story in tech, it has to go back and forth and back and forth.
3:06:29
And the memo that came out on last Wednesday probably sort of reset the narrative there and and and changed how people were thinking about that.
3:06:37
And also, uh these charts are based on momentum.
3:06:40
And so, um it it it's very very hard to say what the >> Yeah, Cloud had gotten to something like uh a million downloads a day according to Mike Krieger. Oh, yeah.
3:06:49
And at the same time, ChatGPT last month, I think, did like 55 million downloads.
3:06:53
So, it's still not outpacing >> Sure.
3:06:57
ChatGPT's lot more downloads >> Plus is number four in the free apps.
3:07:01
Although, how can it be free when you have to subscribe? It's sort of odd.
3:07:05
>> That's a whole thing now.
3:07:06
>> But, bullish for the Ellison Uh Chris says UMF's ruined agency so bad that I'm going to have to start using saying gumption. >> Gumption. Gumption is really good. I like gumption. We hire for gumption. AI gumption. AI agent.
3:07:24
>> hire for gumption here.
3:07:25
>> noun version of gumption? I don't know.
3:07:27
Something Uh friend of the show, I'm Jad, said not having coding experience is becoming an advantage.
3:07:34
And uh Sure Yes says, "Not knowing how to cook is becoming an advantage." CEO of McDonald's. >> I love it.
3:07:40
I'm Jad has some great takes, though.
3:07:42
Uh this is just easy easy riff.
3:07:45
It's it's all in it's all in good fun.
3:07:47
Um What else is going on?
3:07:49
Uh lots of debate over this of the over the singularity pricing as well to Pew put it uh because Barely AI, which I believe is Trung Fan, says Cursors internal analysis shows how hard Anthropic is subsidizing Claude Code.
3:08:05
Last year, a $200 monthly subscription could use $2,000 in compute.
3:08:10
Now, the same $200 a monthly monthly plan can consume 5,000 in compute.
3:08:15
Um there it does feel like there's a capital war.
3:08:18
I'm very interested to dig into the history of of capital wars in in winner-take-all markets, but also in oligopoly markets.
3:08:27
Um we will we'll see what happens. >> on capital wars.
3:08:29
Capital wars are fascinating.
3:08:32
I I lived through it during the Uber and and Lyft era.
3:08:34
There was There was a flippening for a little bit where Lyft was was at the top of the charts and it was a >> the friendly ride-share company. >> It was very friendly.
3:08:43
They the cars would pull up with a pink mustache on them.
3:08:46
They had a very very fun whimsical branding.
3:08:48
And you would sit in the front seat of the car because you were friends with the driver.
3:08:52
>> see Claude doing a pink mustache. >> a real thing.
3:08:55
Yeah, so when Lyft So when Lyft So when Lyft launched, people see Lyft and Uber as like commodities at the at the time, but Lyft was like revolutionary and people were super black-pilled on on Uber because Uber at the time was only black cars.
3:09:08
So they didn't allow anyone to just hop on the network.
3:09:14
Like it was not a three-sided marketplace or whatever, two-sided marketplace.
3:09:18
Like you had to have a like a livery to TCD number, I forget what it's called, but like you had to be registered as a limo driver. And you had a Escalade.
3:09:27
And you normally someone would call you and say, "Hey, there's a wedding and we want you to come shuttle people around at this wedding."
3:09:33
And they would pay you for that.
3:09:34
And Uber went to those folks and they said, "Hey, you are registered to legally drive this car as a professional driver.
3:09:43
We want to book you through our app.
3:09:45
And so the black car became Uber's main product, their only product.
3:09:52
And then Lyft figured out through some interesting, you know, strategy how to allow people to just show up with a Prius or a Toyota Camry or Corolla or whatever.
3:10:03
And so drivers on launch day, I remember like the launch week in San Francisco, would just show up in just some random car.
3:10:11
But Lyft gave them all pink mustaches, huge pink mustaches.
3:10:13
The team can pull up a picture of the huge pink mustaches on the front of the Lyft cars that were very whimsical.
3:10:20
And and it was remarkably cheap because Uber had been somewhat subsidizing the black cars.
3:10:26
They hadn't been taking a huge margin, they were losing money.
3:10:28
But the Lyft cars were way cheaper because someone was just showing up with like their Toyota Corolla or whatever with their pink mustache on it.
3:10:36
And so so you would hop in the front seat because this driver was your friend.
3:10:41
[clears throat] There you go.
3:10:41
The pink mustache on top of the What is that? A CRV or something?
3:10:47
Um And so and so you would hop in the front seat.
3:10:49
You They would ask you to pound the driver.
3:10:52
Give him Give him the knuckles. Give him knuckles. This is a thing. I'm I'm almost sure.
3:10:56
So you'd be like, "What's up?"
3:10:58
And then you'd get in the front seat and they would drive you.
3:10:59
Of course, if you had a group, you'd be in the back seat.
3:11:01
But then >> Everyone had to give knuckles or just >> I I think the knuckles might have been optional but encouraged, recommended cuz it was like, "Hey, we're buddies and we're and we're fighting the war on the Borg.
3:11:10
We're going up against them."
3:11:12
>> encouraged to get in the front seat? >> Front seat.
3:11:13
Front seat in the in the Lyft.
3:11:15
In the Lyft, when you're taking a Lyft, your friendly Lyft driver shows up, you sit in the front seat, you give him knuckles. This is a thing. Pink mustache. That's crazy.
3:11:23
Big chungus sends his regards.
3:11:25
So You got to You got to hire a car, but you have to be actually boys >> You do.
3:11:29
But but truth be told, like a lot of the early Lyft drivers were like super cool, tech forward San Francisco people.
3:11:36
And so like you'd have a lot to talk about because like people were like they they it wasn't professionalized at all yet.
3:11:43
It was very much like like two-sided marketplace.
3:11:45
You'd make a lot of money as a driver.
3:11:47
Like there was no margin compression yet.
3:11:48
It was like free money everywhere for everyone.
3:11:50
And it was a capital war.
3:11:52
And Lyft raised a ton of money and gave and came into the market with a disruptively cheap product and and wound up delivering a really fun experience that was very very cheap.
3:12:01
And then Uber sort of fast followed and launched Uber what's it called? The normal one. Not black. Uber X. Uber X. Yes. Uber X.
3:12:12
And Uber X was bring your own car.
3:12:15
And then and then the rest of the game played out.
3:12:18
And it became a capital fight.
3:12:18
And for a while like the meta was oh you're you're you're you need a lift you need a you need a ride you open Lyft and you check the price and then you open Uber and you check the price.
3:12:28
And people would check both prices constantly.
3:12:29
And then and then over time like the liquidity and and power law stuff came into it.
3:12:33
John, do you remember Uber helicopter?
3:12:35
Yes, that was the partnership with Blade.
3:12:37
I think it partnership with Blade which was the Uber for helicopters.
3:12:40
And it was a thing but in in like you know for a few weeks or something. >> Did you ever do it? No. No.
3:12:47
Cuz Uber Uber's had a whole bunch of different like stunty products like Uber you know Uber yachts or Uber they did Uber ice cream where you could get ice cream delivered.
3:12:59
That was like the precursor to Uber Eats.
3:13:00
They did an ice cream delivery truck where the all they they they every Uber driver checked in with the with the field office got an cooler in the back with a bunch of dry ice and a bunch of ice cream and you could press a button and for like $2 they would bring you an ice cream.
3:13:15
They even did I think Uber kittens and you could ask >> Yeah yeah Uber dogs or something.
3:13:21
You could like I remember that. Yeah it was very funny.
3:13:25
Anyway lots of stunts lots of ways to get attention go viral and that was sort of the uh the launch strategy.
3:13:30
They would do that for whenever they launched new markets.
3:13:33
They would come in and they'd be like we need to go viral on the local news.
3:13:34
We need to let the town know that we've arrived and so uh lots of lots of back and forth in the capital wars of 2023 2024. It was fun times. I lived. I fought through them.
3:13:46
Uh what else [clears throat] is going on?
3:13:49
I think that's a good place to call it.
3:13:51
>> Well, thank you so much for listening.
3:13:54
We will see you tomorrow at 11:00 a. m. sharp.
3:13:56
Leave us five stars on Apple Podcast and Spotify.
3:13:59
Sign up for our newsletter tbpm. com.
3:14:02
And have a great rest of your day.
3:14:04
>> an honor and a privilege and I hope by the time the show goes live tomorrow the war is over. Me too. Me too. I would love that. >> be quite nice. Thank you. Thank you. Goodbye. Cheers.